Close

Project Ideas

eric-hertzEric Hertz wrote 06/15/2015 at 14:32 • 7 min read • Like

I lost all the images here! Links and uploads! 

:(

slowly rebuilding...

Project Ideas... In order of most-recently-listed, first (read from the bottom-up):

12-5-18: YES!

10-24-18:

https://hackaday.com/2018/10/24/av-synth-is-psychedelic-analog-mayhem/

This is kinda where I saw some old ex;periments with LVDS displays going. A shift-reg for the "blue" channel, would position hsync, vsync, and DE, another for the Pixel-Clock, but the rest of the signals ("red" and "green") could be analog... A lot of these displays can handle weird-timings, with equally-weird effects.

CRANKY CHARGER (10-31-17)

(Update 12-12-17: I think the/a flaw with this/my approach has to do with there being no way to *force* the charge to flow from the cell into the battery... with a switched-inductor the charge *has* to flow somewhere... but as an extreme example, an open-circuit cell or capacitor is just as happy holding its charge.)

(Update 12-9-17: @Ted Yapo thinks a car can be started with a coin cell! https://hackaday.com/2017/12/08/coin-cell-challenge-jump-starting-a-car/ )

(Update: 11-20ish-17: I have run simulations... and no luck. Can't wrap my head 'round it... something about capacitors storing *charge* rather than voltage? The 'battery' never charges past its initial value. Huh.)

So... you've got a (small) 12V lead acid battery and a whole bunch of nearly-dead 1.5V batteries, or maybe a single .6V solar cell... and wish to charge that 12V battery...

Imagine that solar cell, or that 1.5V battery as a cup of water, and that 12V battery as a bucket... somehow you've got to raise that cup up over the edge of that bucket, then pour.

So... take a bigass capacitor, tie it across your 11.7V (slightly drained) battery... once the capacitor is charged, disconnect the positive terminal, throw the solar cell or AA in series, and connect its positive terminal to the soon to be 11.71V battery. Repeat. Repeat. Repeat.... no?

Obviously this could be automated and improved... but I'm visualizing it a bit like a crank on a water well, so could just as easily be done with a rotary switch.

I wonder if that tiny solar cell from my old calculator could eventually charge this 1.5AH 12V battery with nary an inductor. Or whether I could "theive" those last joules from my AAA's this way.

@davedarko had a series/parallel switched capacitor-bank a while back... and @starhawk recently reminded me that really-dead 9V batteries are still usable with 4000-series CMOS chips (Unlike four minorly-depleted AAAs running TTL chips, blah!) Bringing back the thoughts of ultimate battery depletion (recycling!).

Oh, of course, the bigass cap will eventually need to be put in parallel with the 12V battery before use...

...And... fsck... not only has the bold button (and others) disappeared, but so have all the pictures, ?!


[SMARTER] IDE Drive-replacement 

11-24-18: This has gotten too long! So I've created a project page. It's still very much in the "idea" stage, though.

https://hackaday.io/project/162439

------

#Let it Snow, starfields, what else...? Maybe a sprite, maybe not... 1k? (12-8-16)

----

Processor-Replacement Revisited (see below) (12-8-16) (UPDATE 2-5-17)

Been looking into 8086's recently, half-tempted to put together an old PC/AT/Compatible for some learning/"hacking" (maybe even develop a system ground-up, but that's unrelated to this).

Occurs to me: 8086's ran at rates from 5MHz - 10MHz... BUT the instructions require multiple-clock-cycles to execute. According to https://en.wikipedia.org/wiki/Instructions_per_second a 5MHz 8086 runs at something like 0.33 MIPS.

Hmmmm....

(see also below: Processor Replacement... (2-4-16))

IT HAS BEGUN: #Improbable AVR -> 8088 substitution for PC/XT

----

mrm, mcp, mmv, meverything! https://hackaday.io/project/10512-potentially-usefulobscure-linux-stuff/log/49872-command-line-utilities-differ-from-bsd-what-to-do (12-4-16)

----

I'll just leave this here... (Thanks @rawe! #"hatching" vectorizer) (3-22-16)

https://hackaday.io/project/9746-hatching-vectorizer/log/34206-line-following-robot/discussion-51095

------------

USB-Drive as a device...? (3-21-16)

@mincepi's #pi2hd44780 has an awesome trick for creating a "device," of sorts, that shows up and is accessible as a USB-Drive on any computer... Now, @usedbytes's #Gadget suggests using an RPi as a "USB-gadget" for various I/O...

I wonder if these two concepts can be combined, showing up as a USB-Drive (accessible on any computer, no drivers necessary)... Making all the "devices" appear as nothing more than files on a FAT32 (for compatibility) drive.

Somewhat doubtful a <PiGadgetMountPoint>/dev/ would be linux-friendly, since /dev is a special folder with special contents, and it wouldn't be particularly Windows/MacOS friendly anyhow... but maybe a universally-accessible "character device" of sorts could be created by, e.g., having a file that's a single byte in length... or, maybe, a "file" that's 2GB in "length." Must re-consult @Paul Stoffregen's https://www.pjrc.com/tech/8051/ide/fat32.html to see how feasible such a hack would be.

(And, who knows, maybe it's not particularly necessary these days, anyhow...? Can a linux ethernet-gadget and/or serial-gadget be driverlessly-accessed on most current OSs?)

---------

Seriously 2.0... (3-18-16)

Every phone needs an infinity-button.

-------------

Processor Replacement... (2-4-16) (and some new thoughts 2-5-16)

Basic gist: Insert an AVR on an old motherboard, in place of the original (e.g. x86) CPU, and use the motherboard as a "dev-kit" of sorts.

Begun! (noted 3-30-17, long after the fact) Check out #Improbable AVR -> 8088 substitution for PC/XT


--------------

Seriously... (1-27-16)

Best I remember, we never see its front, so mightaswell be a smartphone-case...

But here's a dude who's selling one based on the original! Only $599.99!

http://www.replicaprops.com/Quantum-Leap-Handlink_p_62.html

And, appparently, there was another version in the earlier seasons which was more "sleek"... more gray/black... I thought so. hmm...

---------

CD/DVD cartesianThing (9-1-15 2.0)

( Moved to #CD/DVD mechanisms and cartesian thinggie[s?] )


----------



[USB ->] C/H/S IDE (9-1-15)

( moved to #Read The C/H/S IDE Hard Disks (USB?) )

OTHER POTENTIAL USES:

RAM-On-IDE:

Have had a running-idea of creating an ATA-"hard drive" that has a large portion of its space allocated in SDRAM, for the sake of adding large amounts of swap-space on ancient systems... E.G. I have this old 486 Tablet Computer maxed out at something like 16 or 32MB of RAM, and replaced its original 1.8in 80MB hard-disk with a compact-flash card (This was an ORDEAL, the BIOS didn't support anything larger than something tiny like 128MB! WEIRD, for a 486, no? But managed to get it running with drive-overlay software... full 4GB SSD!) Anyways, if I created a drive that was half FLASH and half SDRAM, then "partition" it properly, the SSD wouldn't have issues with being written too many times, but still have the benefit of a (comparatively to the system's design) huge amount of swap-space. (I guess the USB factor is kinda unnecessary, in this particular case, but maybe something like it could be used for e.g. booting Linux or something from a USB device, or an add-on to a normal SSD, and using an FPGA instead... I dunno).

---------

Magnetic Images? (8-8-15)

http://hackaday.com/2015/08/07/hacker-reads-magnetic-stripe-card-with-flatbed-scanner/#comment-2672161

led me to:

So... VHS frames aren't exactly *visible*... probably due to NTSC-encoding (Y-UV, etc...?) and carrier-frequencies... but maybe an image (or series of) could be recorded manually...?

(3-22-16: See Also: http://hackaday.com/2016/03/21/just-when-you-thought-magnets-werent-magic-magnets-are-mechanisms/! There's apparently magnetic visualization strips, now... nevermind printed images.)

---------

AVR-Direct stepper-driving + bipolar-to-unipolar conversion? (6-15-15-2.0)

A convo with @haydn jones suggests it may actually be plausible to direct-drive miniature stepper-motors *directly* off an AVR. They already have protection-diodes that might be fast/strong enough, etc... it's a completely unintuitive concept to me, based on past-experience, but then again, so are miniature stepper-motors... Sure would cut down the devel-requirements.

And an aside, what would happen if a bipolar stepper was wired in a half-unipolar fashion (the two windings tied together at one point)? Maybe driven in a quadrature-fashion...?

Plausible-interesting-experiments.


----------



VFD + LCD = (sectionally-) selectable color filtration (6-15-15)

(Moved to: #(VFD) Panel-filter )

----------















Overview + DIY! (6-15-15)

This is a place for some ideas I've had floating around, as well as, maybe, those that pop briefly into my head... Who knows if I'll ever get around to 'em.

If you do, or even find it inspirational, I'd love to link 'em! If you find something *particularly* interesting-to-do, we might be able to work-out a parts/notes-exchange :)

More... (6-15-15)

That's what's floated 'round the brain and been-written since I first started this page... I think I'll throw newbies at the top, so they show up in the feed easier.

Like

Discussions

usedbytes wrote 03/22/2016 at 09:12 point

> (And, who knows, maybe it's not particularly necessary these days, anyhow...? Can a linux ethernet-gadget and/or serial-gadget be driverlessly-accessed on most current OSs?)

As far as I know, the RNDIS ethernet gadget works driver-free on Linux, Windows and Mac. I would have thought the ACM serial gadget does too.

I'm hoping I can achieve everything on #Gadget over 'ethernet' (driver free). Current thinking is:
 - SSH naturally, needs putty on Windows
 - Telnet - available everywhere
 - Samba file sharing
 - Web interfaces for some tasks - e.g. servo control, AVR programming

Making it show up as some completely custom USB device is also possible (from userspace) with the Linux USB gadget subsystem, which means you can make it do whatever you want and write a custom driver for the host

  Are you sure? yes | no

Eric Hertz wrote 03/22/2016 at 10:41 point

Cool. There's another likely driver-free solution. I dig 'em!

If I understand correctly things like GDB and whatnot often connect via TCP/IP, even if you're connecting to a serial device on the same system... (I know that e.g. OpenOCD works that way). And isn't that kinda the de-facto unix-communication-scheme amongst most same-system applications?

I guess my ignorance is in how to program a user-side interface to it... not being familiar with networking at these low-levels (e.g. in C), and bash-scripts' being so easy to do with redirection to files. OTOH, in my endeavors with openOCD I learned of 'nc' which seems to work well for scripting telnet sessions, so I should probably just get over my afore-mentioned ignorance. ;)

FYI: here's an example of nc connecting to openOCD (which is running on port 4444):

echo "Resetting PIC32"
openocd -f interface/ftdi/myFlyswatter.cfg -c "adapter_khz 400" -f board/pic32_my2xx.cfg > /dev/null 2>&1 &

sleep 1

(
echo reset
sleep 1
echo resume
sleep 1
echo shutdown
) | nc localhost 4444 > /dev/null

(learned from: https://github.com/kinsamanka/PICnc-V2/wiki/OpenOCD-PIC32-Programmer)

  Are you sure? yes | no

PointyOintment wrote 03/20/2016 at 03:46 point

> AVR-Direct stepper-driving […] (6-15-15-2.0)

> A convo with @haydn jones suggests it may actually be plausible to direct-drive miniature stepper-motors *directly* off an AVR.

Baked: https://github.com/clearwater/SwitecX25, #Analog CPU Gauge; I was going to do that in my #Conductor's Pocketwatch too. (BTW, now that I have a smartwatch, I think I'll definitely go the smartwatch app route first for that project, and then build the physical one later.)

  Are you sure? yes | no

Eric Hertz wrote 03/20/2016 at 06:59 point

Hah! Awesome!

After years of transistors and h-bridges, flyback diodes and whatnot... What a mind-blowing concept for me. Kinda reminds me of how strange it's always seemed to me to direct-drive some relays off e.g. 555-timers... Kinda hard to believe magnetism can be strong enough to switch a switch, let alone spin a motor, with only a few milliamps of current and these low voltages.

Aside from those particular steppers, (which are an awesome find, BTW, thanks!) I'm reminded of e.g. the auto-focus system (and plausibly even the zoom-motor?) in an old camera, the likes-of-which might be low-current enough to "direct-drive" as well. Or the "solenoid" used in a wall-clock, obviously, must be dang-low-power to last so long on a single battery.

Or galvanometers, or CD/DVD pick-up heads... so many things that might be low-power enough to actually be driven directly from a couple uC outputs, the outputs themselves being basically a half-bridge (with protection diodes).

Very cool, thanks for the info and the reminder!

The conductor's pocketwatch looks pretty interesting. I'm definitely a fan of mechanical UIs, in this smart-everything era... But I definitely understand the versatility of smart-things can far-outweigh the added weight of yet another thing to carry... Nevermind, if you need a smart-thing as a back-end anyhow. Maybe a wall-clock...?

  Are you sure? yes | no