S.NSA, May 2014 Statement

Friday, 06 June, Year 6 d.Tr. | Author: Mircea Popescu
S.NSA incoming and outgoing
Incoming Outgoing
Description Value Description Value
-- --
Total 0 Total 0
    S.NSA assets
    Account 01.05.2014 Net change 31.05.2014
    Cash 461.89802769 0 461.89802769
    Tangibles 2.87709515 0 2.87709515
    Intangibles and goodwill 7.98496216 0 7.98496216
    Total assets 472.76008500
    S.NSA liabilities
    Account 01.05.2014 Net change 31.05.2014
    Shareholder equity 472.76008500 0 472.76008500
    Total liabilities 472.76008500

      S.NSA has a total of 4`737`075 authorised shares outstanding. The total assets per share implied value is thus 0.00009980 BTC. The cash+tangible assets per share implied value is thus 0.00009811 BTC.

      S.NSA realised no operating revenue this period. The P/E implied value per share is so far 0 BTC.

      S.NSA has Special Stock Warrants outstanding issued as per the IPO agreement, as follows :

      # Fingerprint Shares BTC Par
      1 17215D118B7239507FAFED98B98228A001ABFFC7 3`315`952 331.5952 1
      2 6160E1CAC8A3C52966FD76998A736F0E2FB7B452 1`421`122 142.1122 1
      T 4`737`074 473.7074 1


        For the benefit of inquiring minds that wish to know, I will be publishing selections of the CTOs letter below. As so :

        Dear MP,

        Item no. 1.

        Battery holder was a tough little bitch, but is complete. Two spring contacts - Keystone Electronics Corp. part no. xxxxxxxx. Snugly fits two LR6 cells (series, sum to 3v) or a single CR-V3. [1] We shall have two of these (redundant - like the motorcycle with two petrol tanks.) Landing pads & silkscreens (custom drawing) seen in fig0.png. Tested emprirically with several examples of both cells from different manufacturers (there is ~10% tolerance, apparently, in consumer battery mechanicals. live and learn.) Contact and retainer springs (the former is visible in the pic) have sufficient 'give' to accomodate varying lengths of cell.

        Item no 2.

        Unrouted view of cpu landing pads (fig1.png) - some connections snipped for clarity. IC is a 10x10mm square (package only, pads protrude.) Quartz oscillator (8MHz, PLL multiplied internally) and bypass capacitors (shown) must be in close proximity to cpu, connections for these through vias are undesirable. Manufacturer of my sample board pulled this off (see fig2.png) with a two-layer pcb - by using 0.4mm drills. Presently researching the manufacturability of microvias of 0.4mm drill diameter, at our usual house (gp) and others (we aren't married to gp.) They say they will do it - but question is, what yield (% of dud vias. dud via -> dud board.) Also need to determine minimal annulus supported by GP (the actual, not the claimed.) Would prefer to avoid doing this empirically. See below: The pcb real estate occupied by a single via consists of the drill radius plus the annulus (as discussed previously.) Consult literature for the agonizing details if needed. 4-layer board routes trivially but multiplies pcb manufacture cost by 2.5. [2]

        Presently routing 2-layer board with drill parameters used by 'olimex,' manufacturer of sample unit. If GP yields a sample batch with even one dud, I then know exactly where to take the whole orchestra - Bulgaria.

        One shot at routing the whole shebang presently takes ~2 hours of cpu time ('opteron' circa '07.) Doesn't parallelize, either. [3] One of these days I buy new hardware.


        Steel retainer spring ('C'-shaped) in cover slid into plastic chassis. Connects the two LR6 (AA) if used, for CR-V3 simply provides retaining pressure against contacts.

        fig3.png -> stack-up diagram of typical 4-layer board (to get an idea of why this costs what it does.) Illustration shamelessly cribbed from somewhere. Essentially two ordinary double-layer boards aligned and glued, then plated together. Alignment must be virtually perfect.

        With what? answer: 'eagle' - closed-source linux gizmo by 'cadsoft' (germany) running in (net-less) vm. at one point I patiently re-drew the whole shebang in 'GEDA' - the flagship gnu pcb CAD - which turns out to be fatally buggy - flood fill ground planes break up (!). Wanted to love it, takes S-expressions for circuit netlists, spiffy graphics, open formats, etc. But - how anyone uses it for real work is a mystery. Also tried 'kicad' - router is dumb as a rock. And the GUI snows (!) These round out the list. The fellow who can cough up a serious open source pcb cad package is not yet born, it appears.

        To put this into its just and proper perspective : a project that started in November with an intention to deliver by Xmas is now more than six months late. On the other hand, while the hardware involved is nearing completion at an astounding pace, we've not even had the pleasure of working on the software involved - which may or may not include an infinite number of pitfalls and gotchas. Speaking of which, did you know consumer buttons actually have such a wide geometric tolerance ? Neither did I.

        Fortunately, as the various aspiring currencies sponsored by the various governments continue their inexorable sinking against Bitcoin while the actual capital expenditure of S.NSA is pretty well kept in check we can close on the comforting note that "every investor is making ROI hand over fist in fiat terms", a phrase I've heard somewhere.

        The referenced figures, below :





        Category: S.NSA
        Comments feed : RSS 2.0. Leave your own comment below, or send a trackback.

        One Response

        1. Stan, you're a true talent.

        Add your cents! »
          If this is your first comment, it will wait to be approved. This usually takes a few hours. Subsequent comments are not delayed.