A work in progress, so far I’ve managed to build a BeeCard reader (writing doesn’t work) around an Arduino Mega (ATMega) and a BeeCard purchased from someone who sells old Atari parts (Atari Portfolio.)
I had tried to use the Portfolio BeeCard (32K SRAM) in my Korg Wavestation however it doesn’t provide the battery voltage on a certain pin therefor the keyboard always rejects it as a low battery.
My attempts to write data to the writable BeeCards I have has yet to work. I’m not sure where I’m going wrong, I’ve tinkered with the various control lines and can’t seem to get a write to happen.
My read ability might of been an accident? But it’s what I needed. A sloppy python script dumps the card to a file. Not sure if I will revisit writing the writable cards.
Halloween hit and I wanted to do something cool. The goal was to be able to hit buttons on tablet, and trigger a lightning type effect. I wanted to trigger DMX stobe and LED up-lighting (Color Kinetics) in time with blanking the LED video wall segments and firing off a cue in Pangolin Beyond that contains sequenced static beam shots terminated safely. I also wanted to be able to trigger a few other cues in each program.
Cell DNA allows midi input which is able to switch which cue is playing. However, the laptop I was running it on (that also had the duty of spitting out scraped video over gigabit network to the LED panels) struggled. I didn’t have a ton of time to scale the video down optimally.
FreestylerDMX can talk midi. Beyond can talk midi and OSC directly.
Touch OSC usually targets a single endpoint, so in my case I ended up running the TouchOSC midi endpoint aimed at Beyond, which has a feature of doing midi through. From there I simply shot the output into the rtpMidi utility which then got the commands onto the network as a broadcast protocol and delivered them to both the computer running CellDNA (borrowed from MAGFest) and the one running FreeStylerDMX.
CellDNA was the only thing that had a lot of issues, and that was it drowning in the video codec I believe.
The outcome was a pretty cool looking halloween display that simply wasn’t scary enough. The wind blew away the output from the fog and haze machines often — though some people caught the lasers. And it was cool when kids would straight up ask, “Is that a laser?” Yep.
So this machine is something of a pain in the ass. I decided to pick up a slushie machine before MAGStock, to bring to MAGStock. I didn’t jump on the more expensive Bunn unit and opted for a Elmco unit which was located farther away. The journey to get the machine from Ocean City Maryland was fun, and I visited 5 arcades on the trip. But, so far the repair of the machine hasn’t been the easiest.
For the first post, I’ll talk about the gearbox. In a slushie machine if the slush mix is just water it will freeze. There has to be a certain content of sugar. And when cleaning the machine, or running it it seems like it’s a pretty big risk of it happening. The machine has a optical sensor on the input side of the gearbox so the controller can monitor the RPMs of the motor, but there isn’t an encoder on the output side. If there was one on the output side it could cut power to the motor and save it from expensive damage. But no, they put it on the wrong side.
When I bought the machine, it appears that one side had a cracked gear. The gears are easy to buy, and to be honest removing the gearbox wasn’t too bad! It was made to be serviced, probably because they break easily. So for $60 you can get a new plastic gear and be set!
This fully fixed the machine, until slush mix dripped down into one of the controllers and it died. So that repair is in another post (I’ve cleaned it up but haven’t tested it yet.) Slushie machine ownership is not to be taken lightly, setup and tear down time (and mess) is pretty strong. The bowls sit on a seal and the seal doesn’t always seal so well until everything is cold.
This one is checked off and done, although the pictures don’t show it. This was quite a PITA, honestly.
The place I live had a condensate drain clog. The floor has a drain, but the concrete was graded in a manner that everything flows away from the drain. This led to it getting in carpet, then soaking up into my Galaga cabinet which is made of particle board.
I used Milwax wood hardner and soaked it into the wood really good to stiffen it. I removed the broken soaked bottom panel and replaced it with a cut piece of plywood. It took two tries to get the cuts right, as I measured it wrong on one of the attempts (oops.) Also, given that the wood was flared out a bunch it didn’t go so easily. I added some L brackets, and also lots of pieces of wood glued to the side and bottom. I ordered new leveler feet, although I’m still nervous about having the cabinet sit on the replacement board 24×7. I might just leave it and the new game that is joining it sitting on 2x4s in case the carpet gets soaked again.
Pictures are of the initial removal, I didn’t seem to have any of the replacement in the album and it’s back on it’s feet.
This is mostly a documentation post trying to tie together some thoughts.
A few years ago I worked to reverse engineer schematics and operational theory of the Image Engineering DV decoder for video tapes with digital data of laser shows stored on them. These decoders demux 10 DACs worth of data stored on video tape to drive two color laser projectors.
I have a bunch of tapes, and I would like to archive the content before the tapes degrade.
My original research got me to the point where I found the clock and data coming into the portion of the circuit that demuxes it. But I never put a ton of time into getting a microcontroller to decode it since it requires using an interrupt to detect the break in timing to mark frame start — and once I do get the data moved to a computer I need a solid clocked format to store the content and that seemed problematic. (Tying it to a WAV file as sound card captures audio was one thought.)
Then I picked up a ILDA to ADAT converter from DZ/Ed O’Keefe. They’re awesome people from the laser show hobby/industry that make some nice converters. I was originally using modified ADAT to do this operation but wasn’t happy with signal levels, was out to get their ADAT to ILDA box to give me a nice reference to how things should work when I stumbled across the other version. Thinking more about it, and knowing a little bit about the chip in the box I got to thinking…. if I used parallel to serial shift registers in place of the ADCs I could wire the input side to read the data from the DACs in the decoder. The ADAT system (in my case a HD24) would be sampling the data at 48Khz and the DV playback systems runs slower.
Looking into the wavefront ADAT encoder IC, it doesn’t provide all of the clock signals to clock the ADC chips. I found some info on what could be the circuit needed to do the clock dividing:
https://groupdiy.com/index.php?topic=11498.120
In this case their clock is 44.1khz and I would rather run 48khz. Also, since I need 10 channels then two chips have to be wordclocked together and hopefully the ADAT is happy.
To keep cabling limited between the old board and any new board I was looking at the schematic I drew out in the past. It’s been 3 years and it took a bit to get back into it. Data bus + chip selects into something that can store state into some sorta shift register that is compatible with the data output of the ADCs that are normally used to feed the Wavefront AL1041AG chip:
http://www.wavefrontsemi.com/UserFiles/File/AL_Info/Wavefront%20AL1401AG%20OptoGen%20Data%20Sheet.pdf
A few years ago I picked up a computer w/ a Pangolin QuadMod32 board. This card is an output driver for laser show software from the industry leader, Pangolin. I had a bunch of issues with the card, and managed to work out a deal to buy/swap for a replacement. I own Pangolin’s latest Beyond software which is very powerful and much better so this is strictly a historical endeavor. At SELEM 2016 laser enthusiasts meetup I brought both the luggable Windows 98 machine with LasershowDesigner and an Amiga 500 running the original LSD1000 software. When tearing down I noticed that touching the ribbon cable going to the QM32 board the image was jumping all around. It was obvious that the ribbon cable has issues. On the last digikey order I threw on a DB25 IDC male and female connector, and finally got around to building a replacement cable.
It will get tested next time I break out the laser projector and will report how well it worked. Note, the QM32 is before the ILDA spec so previously I built a converter that takes the QM32 output and converts it to ILDA.
Hopefully one day I will find a QuadMod16 card, which is designed to go in the Commodore Amiga 2000.
UPDATE 8/01/2017
First off, I had a problem in my converter that is a DB-25 to DB-25 that converts the older QM32 output to the ILDA standard. I had the red – signal tied to shutter, which was where most of my issues with output were coming from. After fixing this, my new cable worked 100%.
Also, I now have a QuadMod16 card as well. Will be making a cable for that to go from the card to ILDA.
So the prior Commodore 64 audio issue had seemed to be resolved, but when I went to test it with an actual demo it didn’t seem to work. Frustrated I figured I should look into the 3rd C64 that was given to me by Matt C and Mike W. It came from the surplus of Grande Junction in VaBeach, but had markings on it like it had come from CHKD Thrift in it’s past.
Originally there was no life on powerup, but then that changed. It was spotty and upon poking around the glass fuse holder was to blame. Reflowed solder on it and it will still problematic, cleaned the fuse and holder and it was still spotty, bent holder to get tighter grip on fuse and that cured it.
Next sympton was pink screen, orange screen, but not working. Searching online I found a bunch of pointers to things I had already tried (voltage checks.) Finally pointers said it might be the PLA chip, so I desoldered the one out of the donor board (#2) that supplied the SID chip for #1, and placed it in the socket of #3. Came to life, and there was sound from the demo loaded from the SD2IEC.
The metal shielding in #3 was way rusted, and bottom plastic heavily stained. So I combined some parts from #2, and wirewheel cleaned the top shield (which acts as heatsink for ICs.) I sprayed it with yellow rustoleum paint and it looks like a sad Tonka truck but it will perform it’s function and hopefully not rust more.
So check mark on a C64 finally. Claude and Tom are printing me a case for the SD2IEC — whoot! Now to find more d64 images.
A number of years ago I picked up 4 High-End System Technobeams from Hard Rock Cafe in Atlanta. They needed a little work, so I fixed them up. I am a fan of High End Systems products, and had owned some Intellabeams prior.
A few years went on and my Technobeams started going to a few events. They would usually get tossed in a large truck unprotected, and given their shape are vulnerable to the plastic heads getting damaged. I didn’t want that to happen so I kept an eye on craigslist. Finally some cases (and fixtures) popped up in nearby Delaware, and I picked them up. The 4 technobeams had a proper home. The case configuration is sort of annoying, as the yoke has to be moved in a position fairly different from the way I normally hang them to store them. But to have a case that allows it to sit elsewhere would take a lot more room, so it is what it is.
But the foam. It was this melty sticky nasty probably open-cell smelly stuff. Mildewish, tacky and gross. It had to go.
Over the weekend I finally got a chance (and some help from Supersoaker and Christian and others) and replaced the foam with fresher closed-cell foam. It’s stiffer, a bit of a tight fit but hopefully the smell will be gone forever. This material was recycled from other flight cases, and we used a Harbor Freight hot knife to cut it. I see it as practice for when it’s time to make custom foam to hold all the flat panel TVs for MAGFest.
This one is pretty well documented on the internet, at least on other AMX platforms. AMX is my automation platform of choice, because it’s so damn flexible. You can write anything in the Netlinx language for the most part, and it’s easy to talk to external things be it via tcp or RS-232/485/422.
There is an issue with the NI-3000/2000 series where the serial ports can no longer transmit due to a bad cap. In my case my NI-4100 didn’t seem to have working serial ports, so I reverted to using the NXC-COM2 boards that were in the unit that added an addition 8 serial ports on top of the on board 7.
After struggling a lot with strange issues using the NXC-COM2 ports, I said screw it let’s replace this capacitor. I wasn’t enthusiastic to remove the unit from the rack but really the job went very quickly.
The cap is a simple 35v 10uF IIRC, and it’s originally a SMD cap but way easy to replace with a through hole with bent legs. Look at that pic, it looks nice! Especially compared to dealing with Konami XMen PCB where the caps leaked all over the board and ate the traces.
The board inside the unit is probably the same board used in the NI-3000, just with the additional board at the bottom for adding the expansion boards. If anything, I plan to “downgrade” to a NI-3000 as I like the form factor a bit better, but in the meantime the DB-9 serial port is running strong (still had issues with the NXC-COM2 after cap replacement, I think it’s my wiring.)
So this worked well.
On that note, I will post some stuff I hacked together for grabbing weather from wunderground and getting into the AMX system for pushing weather forecast data to the AMX EnvStat communicating color thermostat. Mine jams up a bit, it’s been frustrating since it falls out of hold modes based on the schedule (I said hold damnit!) But got the weather hitting the stat via a Linux host, and able to poll thermostat from IRC (No clouds in this forecast.)
BK Precision Power Supply test, after chip replacement
Some friends were crashing at my place after MAGFest for the local Awesome Games Done Quick (AGDQ) event. At the time I had a lamp in one of the spare bedrooms but the 2nd spare bedroom had no light source. There is a switched outlet tied to a light switch, but I couldn’t think of any lamp fixtures sitting around that aren’t coherent light or require a DMX-512 protocol. Womp womp, just remembered one in the garage come to think of it. Any how, as a quick fix to solve the issue what better thing to use for a lamp than a 15′ strip of LEDs and a bench power supply? Both were handy, so I grabbed them and cobbled together a 1 minute solution to solve the temporary problem. All is good, except the strip curled back up on itself and shorted. The power supply, a nice BK Precision 1672 with current limiting went into protect mode, but something strange happened. The channel was killed.
Upon investigation the unit would always be stuck in constant current (CC) mode and there was no output. My first suspect was easy, the TO3 transistor might be shorted. No, not it. There were 4 or 5 fuses on each of the power supply boards inside, all were fine. I shot off an email to their support. They kindly replied fairly quickly saying they couldn’t help – but here is the schematics. How is that for service? So awesome.
Looking at the schematics and halfway guessing, I notice a lot of 741 opamps. They are all socketed, so the first easy thing to try is just replace them all. Cheap component, easy work (sort of.) My co-worker Will was placing a Mouser order and offered to let me jump in on it, so some replacement DIP 741’s were had.
Removing the PCB was actually a pretty huge pain. The way the unit is put together, it’s pretty tough to service (although with the bottom of the PCB exposed, I bet shops highly familiar can make easy work of taking measurements.) After freeing the board which requires moving the front knobs, the pot retaining nuts, 4 screws that hold the board in, the metal bracket above the board, and loosening the front plastic and disconnecting a bunch of cables that have little slack I was able to quickly and easily replace all op-amps.
Upon first test it didn’t seem to work right., the voltage adjustment worked like a champ but the current adjustment didn’t — no output. I wanted to keep the unit all together, and figured I’ll just use the right channel and +5vdc.
Upon putting it all back together, I tested it again and it worked fine. I’m thinking one of the connectors wasn’t seated fully, at least that is what I’m telling myself. So the lovely power supply is back in action 100%, thanks for the schematics BK and thanks Will for the 741’s!
