• Ferrofluid generator

    esot.eric02/05/2020 at 23:10 1 comment

    capillary action - via plant illness


    A ferrofluid siphon using capillary action, reservoir at the bottom, a siphon tube U-shaped [upside-down] to a receiving tank higher up; cap-action fills the tube, a magnet is exposed under the receiving tank to draw fluid out of the tube into the higher tank. When full, the magnet is blocked, a valve is opened to a separate pipe between the two tanks, with intermediate turbine.

    Make electricity, profit. copyright me, 2-3-2020

  • Blue LED fades to teal/blue-green?!

    esot.eric01/12/2020 at 02:21 12 comments

    UPDATES At bottom!


    This is a $1 Car->USB-power adapter. When powered, the LED is a piercing blue, but when power is removed, the capacitor discharges, I'm almost certain I see the blue LED CHANGE COLOR as it dims, from a piercing/pure blue to a "warm" blue-green/teal.

    How can this be?! Isn't the wavelength dependent on the material?

    If this LED were supplied with the same average voltage/current, but applied instantaneously via PWM, would it glow blue or teal? Is its color a factor of instantaneous voltage/current, or something to do with "warming up" like a filament bulb? Or is this just an illusion?



    In this second vid we have the blue->greenish LED now exposed without shining throug translucent housing, and another identical device which to my eyes does not green-shift at all. Oddly, the camera [and/or my phone-screen?] shows a distinct greenshift for that one. The original greenshifting one does, still, appear to my eyes to be doing-so.

    I've a vague idea regarding a brief dropout of the switching regulator when power's removed that could cause a brief voltage-spike causing a green-shift that maybe latches until the power drains completely[?], but doesn't increase in brightness since it doesn't recharge the capacitor[?] More on that in comments, below.

    Thanks for the thoughts, y'all! This is intriguing, maybe I'll get more sciencey about it!


    UPDATE 2

    This article is mostly about white LEDs, but mentions wavelength-shift due to PWM duty-cycle [and concluding due to heat]. Could be relevant-ish, plausibly.


    Though, that shift [toward blue, with decreasing duty-cycle/heat] is tiny compared to what I see.

    I tried some experiments with a DVD as a prism of sorts. Hard to capture with the camera, but I'm surprised how wide the "spectral width", as to my eye it clearly looks to cover the whole range from green to blue. There's even yellow and red [though dim]. I really thought LEDs were quite a bit more narrow in their spectra.

    Still pondering.

    Also, still haven't found much regarding peak-wavelength vs. current, *except* in overcurrent conditions.


    esot.eric12/23/2019 at 11:09 0 comments


    I decided I don't like LED lights on my xmas tree, so invested in regular ol' bulbs.... which needs either 120VAC or a lot of rewiring. Since the dash is still open, I tapped off the wiring to my repaired lighter-outlet. I'd, previously, run wire to the back for 12V... but it was powered by a dangling lighter-plug which has since been repurposed, so the wiring's run, but ends were cut.

    It's taken quite a bit of brain-rewiring to get to the point where I'm using 120VAC extension-cords for wiring 12VDC, but, seriously, $1 gets yah a plug, three sockets, and 6ft of wire. Nevermind those plugs actually stay in the sockets [unlike lighter-plugs].

    Not advised. But, functional.

    And, every socket needed gives two spares. I mean how can yah beat it?

    So now that nice warm glow of multicolored incandescents. I'm really fond of that blue. Much nicer, especially in a chilly van.

    Remember when blue T1-3/4 LEDs came out? They were so dim you could barely see them in daylight at 30mA, and $6 a piece when they first appeared in the Mouser catalog! I was so thrilled I bought several, and somehow got so fixated on keeping stocked I never really noticed that the light from blue leds is, frankly, kinda ugly. Maybe it's the narrow spectrum?

    But for xmas lights, there's something special about the blue bulbs. I dunno what it is, they're much darker than the other colors, but the bulb itself glows, and yet somehow that yellowish glow still shines through, but only directly at the filament.

    I dunno, maybe just nostalgia, but lights in winter were always about making long cold nights not only brighter but also warmer.

    I had some "warm white" LEDs on the tree, first. They seemed nice alone [*especially* in comparison to the cool-white], but beside the multicolor incandescents they have a very weird feel, almost ghostly. It feels like they're ejecting dust rather than emitting light.



    And only *just now* it occurs to me I coulda left the inverter up front and actually ran those extension cords at 120VAC. Wow, brain. Wow.

  • Dash STILL open...

    esot.eric12/07/2019 at 19:30 0 comments

    Wouldn'tcha know it, the brand-new heater-fan switch is already failing!

    Yesterday in the full-on position it was running about half-speed, which is actually faster than the other settings, and if I jiggled the knob just-right it would do full-speed.

    Today, no full-speed. Oddly, I'd been thinking it'd be nice to change those resistors such that the distribution was more logical, specifically I wanted a setting close to 50%: e.g. before the new switch-fail, its speed settings seemed around 10%, 15%, 20%, and 100%. 100's too fast, sometimes, but 20's too slow.

    So, was thinking different resistors to make it more like 10%, 20%, 50%, 100%.

    Then, the next day, after thinking this, the new switch[?] started failing, giving me basically 50% where 100% used to be.

    No joke.

    Weird failure.

    50%'s nice, though. So, I'm not particularly complaining, but it leaves some thoughts, aside from the obvious, which I'll ignore for the sake of science. [Especially being that this is easily the third time the heating system's failed-to-/almost/-desired-functionality within a day of my thinking it].

    So, main concern: if it's running at, say, 50% where it was 100%, and 50% is not an option with the available resistors [presumably, the replacement fan motor has a lower winding-resistance than the original], that must mean there's resistance [or voltage-drop] somewhere in the 100% path. This resistance is new, AND jiggling the switch did, for a while, give nearer 100%, as well. So, logically, this resistance must be in the switch. No?

    So, as we found with the last switch, even a tiny resistance in the contacts [as I couldn't tell much difference from 100% fan-blowage] was enough to produce so much heat as to melt the switch's plastic.

    So, then, an even bigger resistance should cause even more heat, no?

    P=VI, V=IR, P=IRI=I^2*R... hmm, maybe.

    I dunno. And no guess as to what'd cause said resistance, anyhow... near-perfectly-distributed oxidation across a sliding/rotating switch-wiper's path? [Or on the wiper itself?]

    Well, through what'd appear to be sheer dumb-luck, I now have that 50% fan-blowage and recirculation without AC running, just as I'd planned on implementing but didn't actually implement.

    What when I want 100%, or AC, again? I dunno. The third case was bad-luck, I'd rather not recall it, though repair turned out to be quite cheaper than I expected, which, after two weeks in the cold, was quite nice "luck" indeed. 50-50, I guess.

  • Dash still open

    esot.eric11/30/2019 at 03:47 0 comments

    fixed the lighter-outlet. Kinda proud-ish of how it turned out, but that'll likely never be seen, and I guess that'd be a good thing, meaning it got the job done. Still, it wound-up easily two full days' work. Two hours' wages [and a few hours searching?] Prb coulda replaced it with a higher quality new one. Or, yahknow, $5 at walmart.

    AC is weird, still with a wire dangling out the hood and through the door to indicate when the pump's on [almost never]. Boggles the mind, defies the manual.

    Next problem with Recirc: it sucks all the heat away from the feet! So, am debating putting a reverse-switch on the fan [50A DPDT?!] Will prb need a [big] relay, instead. And, I dunno if squirrel-cage fans [nevermind the duct-work] will really allow for reverse.

    See, the prb with *not* using recirc is due to exhaust-intake while parked. So, it's kinda important for heat... or get a propane heater after payday. Or go south.

    Radio Aux-in is great, but some-idiot used a short and delicate wire "temporarily." And, it makes a pretty decent 4G antenna in some orientations. So, shielded is the plan, and I've a ton of old inter-pcb wires from old VCRs. oddly, like CD-ROM audio cables, it's all left/right and a single shield. Guess L/R crosstalk is better than [short] ground-loops? Sure, why not. Also, presently, picks up the FM audio signal, a bit. I'm guessing shielded [insteada twisted-pair with ground] might help. Though still debating the best grounding scheme... one end unconnected? Oy! So, "temporary," until who knows when. Oh, also the brilliant idea to use RCA jacks... 'cause there's plenty of space for another huge hole. Gonna be ridiculously-placed, though functional.

    And blower-switch seems OK, I even found a knob in my collection which *exactly* fit... and then lost it.


  • ISA PnP, now that's a hack!

    esot.eric11/06/2019 at 16:33 6 comments

    Meh... wrote this a while-back, realized I got some details wrong, intended to fix 'em, but ah well. The jist's there.


    Always been a running-background-thought in the ol' noggin... how can ISA plug-and-play function on a bus which, by-design, requires each device to have a dedicated address?

    A bus-transaction, simplified, consists of:

    1) output the device-address

    2) set-up the data-bus

       (Hi-Z for read, or data to write)

    3) strobe the /ReadData or /WriteData pin

    Note that the order may vary, slightly, but is important! E.G. once that /WD signal comes through, the data and address *must* already be set-up, and remain for a while thereafter, otherwise you may write the wrong data and/or write the right data at the wrong address[es!]. Also note, there's no "non-address", every address, including 0x0000 and 0xffff [which may result from e.g. *not* explicitly setting an address, which can't actually be done], may have a device or memory attached.


    A) Bus accesses are handled via hardware, with a very specific protocol. So, no bit-banging of certain pins, even if, in a typical access, that pin is only paid attention to while another is active. 

    B) Similarly, pnp devices should work in [older] systems which are not designed for it [the OS will configure the cards]. So, e.g. no turning output-only pins [like /RD, or Address bits] into inputs, or input-only pins like interrupts into outputs, etc. because older hardware just wasn't designed to do that.

    C) There are some other signals, e.g. to allow the processor to release the bus to other "co-processors" such as an FPU or DMA controller. I suppose it might've been plausible to do pin-repurposing by e.g. a slight modification/upgrade to one of these, or even, e.g. releasing the bus to a new device [like an ISA card designed as a PnP-"host"], but, realistically, by the time PnP was becoming a thing, FPUs and DMA controllers were long-since embedded in the CPU and bridge chip. Further, the I/O logic buffers [or, more importantly, input-only or output-only logic buffers] for each pin on the ISA bus would've been in the bridge-chip on newer systems [or dedicated discrete logic on e.g. a PC/XT], between the FPU/DMA/CPU and ISA bus. So, again, changing the direction of a bus-buffer pin which was designed to be one-way [e.g. /RD, or INT] would not be possible even when the CPU releases the bus. So a theoretical "PnP-host" add-on card still wouldn't be able to e.g. turn /RD at the bridge into an input, nor, maybe, even Hi-Z. Thus, it'd still be driven by the bridge, even while the hypothetical PnP-host card attempts to do the same even, again, if the CPU were to release its control of the bus.

    So, basically what I'm getting at is that the only way to access an ISA card is via a well-defined protocol which basically precludes any sort of pin-repurposing, bit-banging, etc.

    [I'm sure there are more challenges, but I'm on a roll]

    So, then, how can one configure a PnP device whose address has yet to be configured? How do you tell it you're talking to it?

    Sure, it could just *listen* to every transaction, to/from every address, but how can it know [how/when to] respond or act upon whatever it "hears"?

    And, further, how could this work in a system with *several* such devices. How can one be configured differently from the next? They don't even have the ability to respond, certainly can't talk directly to each other, and the system doesn't even know they're there!

    And, further-still, in a system with *several more* non-PnP [nor even PnP-aware] devices which could literally be located at *any* address, and no defined protocol for identifying themselves, how are we to avoid conflict *during* configuration?

    So, then, one might think "well, obviously, the first step is to at least get their attention," and that means some sort of signal sent via normal ISA bus transaction... and... whatever signal that may be, it can't cause havoc with other devices... ...

    Read more »

  • While Dash Open

    esot.eric11/03/2019 at 10:53 0 comments

    From the last log, it started with a burnt heater-fan speed switch, yeah, it was gnarly.

    The housing around the contacts/terminals crumbled out while exploring the damage.

    I had bought that [the third] two winters ago, a cheap push-pull SPST on/off 20A switch, rather than the 4-position rotary-switch designed for the job. I wired it up so Off was the lowest speed-setting, which places three resistors in series with the motor and ground. On, then, was the highest speed, bypassing those resistors, wiring the motor straight to ground.

    I almost, this time, bought a slightly higher-quality SPDT toggle-switch, 35A this time, which'd'a given three speeds... but, it would still be switching directly between essentially-off [low-speed] and full-speed... inductive arcing, much? 

    So I started thinking. Per chance, a few logs back i'd done some research on switch-contact-ratings, for another reason entirely, and one of the resounding factors in such ratings is *inductive loads*... like motors, which can cause arcing during switching that'll destroy the contacts' conductivity... so much a concern that, say, a 5A 120VAC switch is downrated to 5A at 14VDC (fourteen!). It's kinda a big concern! 

    Now, I dun have with me an ammeter capable of measuring the current-draw of my blower-motor... its fuse is 50A, and it just destroyed a 20A switch. And that after what looked to be a slightly-higher-quality-but-similar switch a couple years back, and *that* was a replacement for [I assume] the original factory-installed switch. Yesterday's purchase is *at least* the fourth.

    So, more thinking... yahknow, they could reduce quite a bit of contact arcing, in this shorted-series-resistor arrangement, by using make-before-break switching... and that'd be comparatively trivial to do in... A Rotary Switch. [Like the original].

    So, I coughed up the extra $17 for an aftermarket switch intended to replace the switch designed for the task, on a whim that said design mighta considered make-before-break, the intended current-draw, and likely more I hadn't'a thoughta... *and* on the whim that said aftermarketers followed-suit...

    Well, sure-enough they did. It's actually more elegantly-simple, yet more fully-functional, than make-before-break; because we have tap-points at each resistor, they merely short the resistors that would otherwise be bypassed to ground. 

    So, like, the lowest speed-setting no contacts are shorted, all the resistors are in series to ground. The next speed-setting shorts the first tap-point to ground through the switch... But, of course, the resistors are near the motor *and* grounded there, and the switch and its ground are several feet away, so it's more like a small resistance in parallel to that first [small-resistance] resistor. Now the higher current through the motor has two paths, reducing the strain on the switch-contacts [and wires].

    Then the same for the next speed-setting, now there are three paths for the even higher current.

    It's a strange blend of cheap, lazy, and yet highly-effective in a way that regular "high quality" general-purpose switch-designs would be hard-pressed to mimic. 

    Take it a step further, the actual process of physically engaging contacts can wear them [and the bearing-surfaces, etc.] out, especially considering the surface-area necessary for such loads, but in this design each contact only makes/breaks once over the entire range!

    Definitely worth the extra $17 for the learning-experience, alone. I can't vouch for the quality of this particular implementation, yet... Before I had the make-before-break idea, I was pretty turned-off at its being housed in plastic, being that overheated plastic was the problem at least twice before... OTOH, with the potential for much-larger contact-area, and the seeming solid single piece of formed metal all the way from the connectors to the wiper [rather'n, say, a connector riveted to the contacts], there's not nearly as much room for resistance...

    Read more »

  • AUX in!

    esot.eric10/31/2019 at 08:36 0 comments

    Heater fan switch has been getting so hot as to fill the place with burnt-electronic-smell... it was probably bound to happen, the biggest switch I could find that day was 20A and the blower-fuse is 50(!)A. Also, inductive-loads, arcing, etc...

    So ripped out the dashboard-cover at some late hour, and leaving it 'till payday for a better switch. 

    [Note to self: heat-selector-switch controls blower-relay, speed-switch *directly* switches in and out series resistors with plausibly 30A blower motor. This is the third failed switch. Maybe turn heat *off* before adjusting speed.]

    [Also, obviously, thoughts on relays or even PWM.]


    Meanwhile, a few long-running project-ideas for while the dash is open...

    Hacked the cheap FM/AM-only stereo for AUX-in, can finally play my own tunes [and GPS directions]! [Yes, I have an FM transmitter. It sucks, and my phone's *really low* max volume made it far worse.]

    Stereo was a bit quirky, I think it's Class-D, and apparently that means the two identical large-ish caps near the [speaker] connections *aren't* for the speakers. None of the chips have datasheets, wiring is crazy [rear speakers not implemented, so they ran front-right - signal trace to rear-right +, first, shorted that with rear-right -, then fed that to front-right - via two parallel 0-ohm resistors, and such-like. 

    Ultimately found two parallel traces with large series surface-mount caps originating under the tuner-circuitry's shield heading to unknown chip which is connected to both the presumed processor and the presumed amp, presumably for things like equalization.

    Oddly, it has four parallel traces leaving it through four of the same series caps, to the amp. Prb also handles r/l and f/r balance... if f/r was implemented. But why bother with wiring caps to unused outputs? Or maybe it's differential, hmmmm... plausibly isolating amp power/gnd from the rest... hmmm...

    Also oddly, amp looks identical [sans different markings] to the other chip with the big caps... am thinking it's for power.

    So, it was a crap-shoot, but I tombstoned the caps and wired up the phone and, yep! 

    Managed *just enough* space for a dpdt switch [on-on-on, weeeeeeee!], and dangling a headphone plug, 'till I'm up near storage again.

    LOTS of debate about how to shield... ultimately the sun was going down, losing light to work by, so decided on two twisted-pairs of wire for each r/l signal pair. All "shields" tied together at the chip's ground, but at the switch they're N/C.

    The aux input is split kinda like a mini-plug to rca cable, each signal "shielded" with a twist, then the shields tied together at both ends [one at mini-plug, one at the switch] then a single ground-wire down to that same point on the chip. Trying to prevent r/l crosstalk...

    But, sure-enough, with no source connected to aux, and volume-maxed it definitely picks up the radio... not sure if it's a shielding issue, or inside the switch, or...

    But *WAAAAY* better than that FM transmitter, which was too quiet, staticky, and picked-up ignition-firing, and even worse when the phone was charging.

    On that note, not sure what to do about charging simultaneously... definite ground-loop-ish issue, seems to pick up the dc converter's switching[?]


    Other while-dash-open ideas include:

    12V banana plugs, ferchrissakes.

    [Maybe 5V and-or built-in USB? + switch!]

    Heater only does recirc with AC running, by stupid design. Add AC cut-off switch

    Aux Lighter-outlet melted-to-short. replace!

    Surely more... blanking.

    Oh, stereo power-source selector-switch!

    Also might look into cleaning duct-work.

    Still... surely more...?

  • toggle switch VA ratings, and On-On-On

    esot.eric10/25/2019 at 19:22 3 comments

    Have an interesting toggle-switch... pretty high quality, made in the USA, wound-up in my collection probably more than 20 years ago [I'm pretty sure I remember the day/person], and... it's still available new at regular suppliers. [This is the kinda quality design we're talking about]. 

    C & K 7211.

    Two odd things I never noticed about it in my previous uses, guess it didn't matter there:

    A) On-On-On (DPDT, three positions, so what's with three Ons?!)

    B) [Relevant to current project] "0.4VA MAX"

    First-off, if VA means what I think it means, then this switch is no-go for my project. [And sure-enough, it would seem that way].

    I had a bit of difficulty finding a decent explanation of this rating, so here's the best reference I found:


    Briefly: there are two common contact platings, gold and silver. Gold's used for low-power [logic], silver for high-power. 

    Silver oxidizes, so arcs caused by high-power switching burns off that oxidation. It's by-design! So silver-contacts aren't so great for low-power/non-arcing switching. 

    Meanwhile, such arcing would burn off the gold-plating on low-power contacts, and you'd be left with tin or copper [or something] contacts which make poorer contact than silver, and also oxidize and such.

    So there you have it.

    And, from the sounds of things, 0.4 VA MAX at <28V [20V AC or DC, per my datasheet] is pretty much standard for gold contacts...

    And, yes, that does seem to be a measly 20mA at 20V = 0.4V*A = 0.4W. But it's a little more complicated than Watts, what with AC, RMS, peaks, etc. Or with DC and inductive loads [motors, or even long wires] and Capacitive loads [with large inrush currents, *during* switch-bounce!].

    So, I guess the answer for my project is that, sure, this switch, if it had silver contacts, would handle 5A @250V [AC]... I thought it seemed pretty heavy-duty... And it'd probably do the job of switching 40VDC unlikely-max 3A [most-often a fraction] for a while even with bare copper or tin.. but probably not an ideal solution.

    BTW look into that document, also, for info on derating for DC circuits [no zero-crossing = longer arcs] it's something like x Amps @ 120VAC -> x Amps @ 14VDC!


    Oh, and finally, On-On-On, that's also nicely-covered in there. Basically it's like a typical three-position DPDT switch, but the middle position puts it in a weird-to-me state where the two poles are in opposite positions [What?!].

    I once used these guys to drive/reverse the motors in a wired toy-car [hahaha 0.4VA my a**]. It worked... but now I see the middle-on was actually *braking* the motor [shorting its winding via one of the power rails]. Hadn't noticed. But, could be handy.

    And, finally, they can be wired as a SP3T switch... cool!

    which, also, makes for Off-On-On as a possibility [e.g. Off, Slow, Fast, or Off, Dim, Bright]

    Whoda-thunk? Certainly wouldn'ta thought to look for such a switch, expecting it to be somewhat common!

  • Low voltage power distribution pt.2

    esot.eric10/24/2019 at 22:42 3 comments

    OK, I think the conclusion is the renowned star-topology power-distribution idea, oft recommended in PCB design and system cabinets, is not really such a great idea if you've a floor-to-ceiling system cabinet and attempt to run power wires neatly.

    Or, in this particular case, also not a great idea.

    Which brings us back to the fancy thing [aka main selling-point, to me] about the DC converters I'm using: they've got isolated-outputs. 


    A) all the voltage-converters can be powered by a single voltage source [a largish/heavyish bank of series batteries]

    B) they can be located anywhere in the system [I'd like to keep the power-system as a single self-contained unit in the overall system. In part because this power-system is multi-purpose; this *particular* setup/use just happens to be the most complex]

    C) as long as I run *one* device [rather, "endpoint"] off each converter, there's no concern of "ground" [or other power] "loops" [which, as detailed previously, can be quite a bit more gnarly than the audio-nuts' much-loathed "hum"]

    The last point is key to the latest ramblings.

    I'll drive the computer off a 15V supply, with a local 5V converter [am actually thinking two; a separate one for USB-powered devices, even though one 5V 6.8A converter should be plenty for both]. [These from scrapped car-usb-chargers rated for both 12 and 24V vehicle systems]. The computer is somewhat mobile, thus the 10ft power cable.

    Then hard drives [e.g. backup, large/mass-storage, music, software installers, etc.] can be connected as-needed, minimizing cabling and power-consumption. Each of my IDE-to-USB converter-systems ultimately require 12V going directly to the drive [thus it should be pretty close to 12V, and somewhat stable]. Some have internal 5V converters, others do not. One has a 120V AC power-supply, that will be removed. I'll throw 12V-to-5V car-adapters in, as-needed. Thus, the drives will run off an *isolated* 12V output, again preventing drive/load-current from travelling/"looping" through the USB cables to/from the computer. Then, I'll do the "star" to these guys, as-needed, but keeping them close to the power-bank.

    [Which suggests a wise choice may be a long-cord USB hub nearby... to-ponder].

    I think that'll do it. There may be some looping between the 12V supply and the various drives through *two* USB cables... I guess these things we just don't worry about. [Seems odd to me]. Again, say one drive is sleeping, the other full-tilt, ~0A and 2A, say the power cables are two feet , even with beefy wire that could be 0.2V difference between "common" at one and at the other, now there's a third path between 'em through the USB cables, the system won't be happy with 0.2V between two different points in the "common" [nor 5V!!! COME BACK TO THIS] circuit, thus the running drive's load-current will actually be split between its main power cable and through the USB cables back down the sleeping drive's power cable. This still seems bad to me, and again, this setup is now hardly different than the ol' two-drive IDE setup in a full-tower PC. Weird.

    "Come back to this": we have *another* problem, potentially, in the 5V system... were the drives' 5V coming from the same source endpoint [like common], we'd have the same issues with 5V as with common, since USB also carries 5V.

    But, no, it may be even worse, now, as we have two separate converters essentially *fighting* each other *through* the USB cable!

     Imagine one is trying to output 5.01V, the other 5.02V. And, again, imagine one drive runs at 2A, the other sleeps at ~0A... Seems *likely* one converter, plausibly the far one, will take the brunt of the load. Also seems plausible, in the opposite case, one converter will *sink* current, which most are not designed for.

    ...So, what...? Cut the 5V path in the USB system, maybe...? Fine, maybe even some hubs have diodes on the 5V line... Alternatively, run all drives off a single 5V converter [not ideal, more wires,...

    Read more »