Vagrearg Logo  
Executive Decision Maker

Making decisions is a hard thing to do, so it seems to most of us. But we see plenty of examples where this does not apply. I must admit that randomness is key at the executive level (and I can tell from observational experience). Actually, executives love to postpone making decisions as much as they can. Only when forced, they will actually decide. The typical PHB reflex is about as good as you can imagine a random generator. With some effort and mental ability you can exert influence, but you never fully understand the result.

Anyway, while looking through boxes at the local makerspace for the Pointless Switch Machine, I noticed a few bags filled with LEDs. They are 3rd choice quality, but there are many of them, really many. They must be used for a project. Also, the NEXT conference is nearing, where we are invited, and we need to show off. With that in mind and the desire to make decisions a lot easier, I just had to make the Executive Decision Maker. And even though I made this to make decisions, you can also use it to confirm your believes after the fact. Just ask yourself a question and press that button.

It would have been a lot easier to use a small microcontroller, but then again, there is pride in using stuff nobody needs anymore. I almost used a PIC10F200 when I failed to find suitable TTL chips, but then I found a box with old 4000-series cmos chips, and there were plenty of options again.

edm-pcb edm-pcb-back
The Decide-button selects the appropriate YES or NO, depending on the question in mind at the moment of the button-press. The Executive Decision Maker automatically tunes into the aether and the subconscious of the user. Any answer will, in all honesty, be exactly right as derived from the state of the universe at the instant of button activation.

Here I asked myself the questions:
"Was it right to make this device?". The answer was clear: "Did I use enough LEDs?". The answer was expected (the bags with LEDs are still full):
edm-pcb-yes edm-pcb-no

As seen by the above examples, the answers are right. The answer will clear after 4 seconds, or when the Fast-button is pressed. You can make decisions faster when holding the Fast-button down (see video below).

Some comments on the schematic:

  • The LEDs are of such (bad) quality that the voltage drop varies between 1.95V and 2.25V (remember, 3rd choice). Both intensity and color have a significant spread too. This makes life more difficult, but it was determined that a constant current would give nearly the same intensity for most of the LEDs. All LEDs were then sorted for color/intensity and a current source implemented (~25mA).
  • The number of LEDs in a chain was adjusted a bit after I started soldering them. It turned out that the initial seating plan did not fit and some LEDs were omitted (red went from 51 to 43 and yellow from 63 to 55). However, I didn't bother to do re-ordering of the chains to balance them. The current sources take care of the problems for me (as long as there are at least 3 in a chain, due to power limitation of the transistor).
  • The Intentional Synchronizer only works for true believers. Sceptics will simply see a random answer. True believers, however, are universally coupled and see things differently (they use google to confirm that). Anyone who can prove a statistical abnormality from chance(see notes) is encouraged to contact JREF.
(source or pdf)

Making a box...
Three hours later...
The transformer was scavenged from an old CD player which had its electronics already removed. The IEC connector was found somewhere on a shelf alongside an opened up ATX power supply that had seen the best of days.

Questions, so many questions. I need answers!

"Is it a nice box?": "Was it difficult to make?":
edm-box-yes edm-box-no

And we have a video.

And I finished the box by adding the Intentional Synchronizer (as seen in the last part of the video).


  1. The chance of yes/no is not exactly 50/50. This is due to differences in TPHL/TPLH and TTHL/TTLH. The datasheet says that TxHL and TxLH times are the same, but, in practice, they are not. Even minute differences will skew the result. Is this a practical limitation? No, the differences are so small that we'd see a deviation of <10-3. At a clock-rate of 630kHz we'd see a skew of max. 0.0001575 per ns. You should also note that the flip-flops are configured as divide-by-two. This ensures maximum possible symmetry.
  2. If, for some quirk, you manage to synchronize your body, mind and finger with the clock, you will be able to get consistent results. Just make sure you fine-tune your brain to ask the right question and ascertain you know the initial state.

Posted: 2011-08-11
Updated: 2011-08-14

Overengineering @ request