Project Logs

Collapse

• Test Example Post: Mainframes behind the Iron Curtain

  [skaarj] • 12/03/2016 at 17:09 • 0 comments

  December 3rd, 2016 by [skaarj]
  

  We've covered a very small number of topics about computing technology behind the Iron Curtain, so here is a fresh hot one: [skaarj] is rebuilding a Romanian Mainframe (translated link). Such pieces of communist technology are pretty rare and most of them require excruciating painful hard work to resuscitate. [skaarj] is not entirely sure which mainframe type to chose from the three dinosaurs roaring in Romania's computing centers during those dark prehistoric times: FELIX, CORAL or INDEPENDENT, as the components are pretty rare, in very bad shape and whatever parts he managed to salvage... come from all three types. Due to their proven performances, those three names were a shock for the COMECON market controlled by the Soviet Union, so Big Uncle Bear officially restricted any exports through special directives straight from Moscow. Some government sources explained the decision to Romania's beloved leader due to CORAL and INDEPENDENT being far superior to the USA's PDP.

  So far [skaarj] managed to recover three racks from a Felix-type mainframe, some parts from a CORAL 4030 mainframe, a broken tape drive from an Independent i-100, a Siemens Sinumerik punched paper tape reader from a Felix-M18 and some various peripherals such as 8" floppy drives, huge hard disks, cassette tape drives or various circuit boards. The clumsy serial terminal called DAF-2020 (Display Alpha Numeric) is already up and running, and through the courtesy of [Nico] the future mainframe will be able to access PERTEC-interfaced peripherals through a PERTEC to SCSI converter.
  

  Since the central processing unit (drawer) released its magic smoke in the final agonizing days of the communist regime while running a modified smuggled copy of RSX-11 Operating System, [skaarj] decided to hook a Marvell OpenRD Ultimate Development Platform and install FreeBSD Unix to get access to the functional peripherals and to experiment with ferite core memory banks. Check the videos after the break.

  
  

  Siemens re-branded StorageTek Industrial (PERTEC) interface tape drive
  
  

  The Romanian links are provided through Google Translator, as Language of Dracula slang is pretty hard to understand. Unfortunately, the Romanian Electronics Forum administrators changed the policies at the beginning of this week, requiring an account to view the attached pictures. [banner picture source: West Romania Museum-Muzeul Banatului, Timisoara]

  We've covered a few topics about Iron Curtain Home Computers and Peripherals before, and we've also learned about those dark times from one of our most respectable hackers.

• Project status: scrapped

  [skaarj] • 11/03/2016 at 19:07 • 1 comment

  The project has fulfilled its purposes but it did not caught the desired attention.

  The automobile has a respectable age and it is a hard work to keep it running according to the security standards of the day.

  In conclusion, probably at the beginning of the next year when the insurance expires, the whole system - automobile + accessories - will go straight to the scrap yard for material recycling. At least I will be able to recover a part of the investment spent so far to prove two very uninteresting subjects which were solved in my under-financed project:

  1. Pollution emissions can be reduced using a simple approach, adaptable to most of the ignition and diesel engines;

  2. Mechanical faults in automobiles can be prevented by means of artificial intelligence approach.
  

  I will be back with pictures from the scrapping/crushing/recycling process - if I will not forget the camera inside the car.
  

• Contest entry video uploaded

  [skaarj] • 10/09/2016 at 21:50 • 0 comments

  
  

  Made in a hurry to cover most of the topics presented in this project.

  Later edit - enhanced quality + bonus - added night vision video

• Assistive Technologies - Building cheap night-time vision goggles

  [skaarj] • 09/14/2016 at 06:24 • 5 comments

  First of all, we need two specialized vacuum tubes for amplifying the ambient light. Again, the Soviet Union legacy proves to be useful: V-8 tube.

  B-8 (Cyrillic "B" in English it's V), made in May (5th month), year is 1981. There's also the manufacturer logo, probably some military factory, and also the manufacturer-issued serial number.
  

  In front it's the light receiving window.

  In the back side there's a phosphorus window which shows the intensified result.

  This is a 2nd generation light intensifier tube, one of the best available. We need two such devices, a toy binoculars and a helmet to install everything.

  This tube has an anode (between the fingers) and a cathode which is located as a metal ring in the exterior. It must be powered at around 15kVolts to get the desired results.

  Binoculars:

  Somehow I need to replace the prisms with the tubes, and this baby has a lot of space.

  Left side tube is already installed. These details cover the right side install.

  Protection ring comes with the tube - if acquired from military surplus source. Need to make 4 (four) holes, 2 millimeters in diameter. Phosphorus window is oriented to the eye lens.

  A little fine tuning of the lens distances will be performed after powering on the system.

  In order to accomplish this, we must ask for help from the Soviet Union military surplus: two high voltage rectifying diodes and a transformer-based voltage converter, in order to power it from 12V (car lighter):

  This is a classic single transistor, single transformer feedback oscillator.

  Please excuse the dirt on my hands and on my desk. This is my work place laboratory, I'm currently performing some maintenance for some oil drilling pressure sensors. The dirt I handle in the oilfield is cleaner than the dirt they handle in the offices and I'm proud of that.

  You are probably wondering why do I prefer this ancient converter instead of building a modern one with some microcontroller, a PWM output pin, some FET transistors and a small switching transformer.

  First of all, this is insulated and protected from shock and moisture, including these two high voltage rectifying diodes.

  Second, I need this baby to balance the binoculars weight on the helmet: binoculars stay in front, converter in the back side.

  I'm wearing the helmet while driving, this is stressful enough and I need the hands free of any additional tasks.

  Helmet is also military surplus, used by tank crew in the old days. Keeps warm, it has earphones included and it can be "upgraded" with metal work.

  Helmet is balanced...

  ...and ready for final optic adjustments.

  Night-time test drive:

  This is raw video footage recorded with a Samsung Galaxy Trend Plus S7580 rear camera. It reacts weird to bright images, normally the picture is clear. Anyway, when brightness is low, the camera can capture/record the road and trees in clear.

  Message for the feds/police: THE FOOTAGE IS RUNNING AT 2x PLAYBACK SPEED, I DID NOT DRIVE LIKE CRAZY!!!

  In the mean time - if you decide to build this thing, you must be careful not to power it up in day light without covering the lens. You will burn the phosphorus windows and the tubes will become useless.
  
  

• Assistive Technologies - Let's improve night time driving

  [skaarj] • 08/23/2016 at 07:34 • 2 comments

  Allright, I'll keep this short for now because I'm at work and boss will get very angry if I'm discovered writing a new log instead of watching the monitors while the antivirus is doing its job over the entire network.

  As my colleagues from hack-chat already know, I went to Greece taking my princess for a test drive in May this year. I left Romania, I passed Bulgaria during night time and I arrived at Makaza-Komotini pass in the morning. Around 8 hours of driving in pitch dark in the Bulgarian mountains. Because I read all sort of weird news and reports in the newspapers, I decided to build (hack) a counter-measure to get there safe and also to get back in one piece. There are some gangs in Bulgaria which simulate a car accident or car malfunction, and ask people to stop and provide help. Basically you're driving, it's night, sometimes fog is present and there's this guy with a ramped car waving his hands desperately. You, a decent unsuspecting driver, decide to stop, come out of the car and offer to help. Five seconds later some more guys appear from a nearby forest with Kalashnikov guns pointed at your head. From this moment your life is on the table, Mafia is the poker dealer and the house has all lucky hands: you lose your car, your money, your phone, your daughter, your wife, your clothes, even your underwear. The ramp-jacked car also leaves and if you are lucky, they let you barely alive. Broken bones are mandatory.

  How can you avoid such incidents?

  You need some help provided by Assistive Technology. You need to see them from far away. You need to see the road as clear as possible to avoid any other nasty stuff such as pieces of wood with nails or spike strips to blow all your tires. You don't need to win the round. You need to survive and to keep your family safe.

  When I returned from Greece one week later, I saw some guy with a stopped car and the spare wheel near it. He waved his hands at me. Behind some trees, my googles informed me there were more people, so I kept going and informed the Police.
  

  I will be back with building details.

  In the mean time follow my advice: don't waste your sleep() time. It does bad things to your heart.

  [END OF TRANSMISSION]
  

• Automation: let's build this!

  [skaarj] • 07/15/2016 at 04:40 • 4 comments

  Unable to comply.

  The hard work done until this time was not recognized and not acknowledged so far [not even as blog articles] and will never be. There are other projects artificially enforced to be liked and followed. Small steps on the hard road of science - as an example: trying to reduce lung cancer - are considered to be less important and are better ignored - enforcing political correctness tactic. Occult worship such as Pokemon Go is the important subject these days.

  Fatal error: finance reserves below zero.
  

  System halted.

  ______________________________________

  kernel panic: improbability coefficient below zero
  

• Artificial Neural Network-based Mechanical Faults Diagnosis (III) - Vibration detection test

  [skaarj] • 07/08/2016 at 10:25 • 0 comments

  Greetings, fellow hackers.

  Big news: the hardware ANN speech recognition chip was fried during one of the many storms sweeping my area. No more chances to get another one in time for the round finish. Previous work with denied access to normal equipment lead to reviving a piece of junk to perform exhaust measurements with strange results, including magic smoke. ANN is now fried. This means I'm already disqualified from the Citizen Scientist round, no need to wait for the results.

  This chip can be found at Images Scientific Instruments in the US. It's expansive and customs taxes to bring it to Europe makes it unaffordable for the moment. Also, even if it arrives here the next day, the customs will keep the package for another one to two weeks due to bureaucracy. No chance to get experimental results in time for round finish.

  It worked the first time and the experimental results were already presented here as a scientific academic paper. It should work in this experiment without any problem. We'll see. My [WWK] (without wife's knowledge) funds are depleted, no chance to get another one. Under current conditions, not this year for sure. But I don't worry because nobody cares about that anyway.

  This system can be successfully used to acquire data for a certain type of automobile to identify bad mechanical components by deliberately inserting malfunctioning parts in the wheel axles and also engine block (to get the desired training data) and learn from these situations to detect similar conditions. It can be used in automotive services to easily identify problems - or at least have a clue where the problem may be located. It can also be used by manufacturers who wish to study more about the behavior of their products. Or to understand the purring of your cat, the barking of your dogs, or whatever vibration-pattern-recognition idea you have in mind.
  

  I only have the video in the bottom of the page - live vibration monitoring with baudline on the laptop screen.

  Later in the video when the car is moving, you will see the ANN system boards. There are three leds showing activity:

  - blue led and red led which alternate - located on the input switching board. When blue led is on, ANN chip input is connected to ADXL accelerometer. When red led is on, ANN chip input is connected to GND;

  - blue led in the right side, near HM2007 ANN speech recognition chip: blinking confirmation light showing the ANN detected and processed the vibrations.

  Numeric display shows some signs of instability - there's some bad soldered connection somewhere under the mainboard.

  Thank you @ThunderSqueak for going through your electronics cache to find a spare HM2007.
  

  VIdeo embedded below.

  [END OF TRANSMISSION]

  ______________________________________

  kernel panic: improbability coefficient below zero
  

• Artificial Neural Network-based Mechanical Faults Diagnosis (II)

  [skaarj] • 06/26/2016 at 09:02 • 0 comments

  A.K.A. the "Blonde Power"

  As presented in Science time (I): Artificial Neural Network-based mechanical faults diagnosis (I), this system is based on a hardware implementation of artificial neural network encapsulated into an IC called HM2007 which is used primarily for spoken words (speech) recognition.

  Searching the internet for schematics based on this circuit will show the classic design and two PDFs - the datasheet and the errata specifying a bug design and a workaround which does not completely give the desired results.

  First, the block schematic:

  The following modules are present on the block schematic:

  - ADXL accelerometer

  monitors the vibrations on the wheel shaft and transmits all three connected X, Y and Z axes signals as an output;
  

  - relay board for input selection

  According to the HM2007 datasheet and the jumper settings on the mainboard, the "listening" time is 0.9 seconds. Less than that - ANN says "word too short". More than that and ANN outputs the error "word too long". So the ANN input is switched to ADXL output for 0.9 seconds, and for some 0.3 seconds it is switched to GND.

  Contains a 12V relay and some signaling LEDs: blue for "Vibra Input" selection and red for "GND" selection.

  "STAB" comes from "stabilizer" and it means "power regulator". Turns 12V to 3.3V needed by ADXL335.
  

  Relay handling is done by the PIC16F84 through an ULN2803A open collector circuit.

  Connector in left side:

  PIN	DIRECTION	FUNCTION
  1	INPUT	+5V (extra confirmation)
  2	INPUT	+12V (ULN+relay+STAB PWR)
  3	OUTPUT	+3.3V (PWR ADXL335)
  4	OUTPUT	HM2007 VIBRA IN
  5	INPUT	ADXL335 SIGNAL OUTPUT (X+Y+Z)
  6		GND
  7	INPUT	PIC RA1 RELAY CMD
  8		GND

  Before any comments such as "why wasting an entire ULN2803 circuit when you could just use a transistor", please understand that I worked using whatever components I can find in the drawers or recover from malfunctioned devices I find in sometimes abandoned warehouses at work.
  

  - Decimal Numeric Address Display board
  

  Displays the error signals or SRAM addresses issued by the HM2007 when certain conditions are met.
  

  - mainboard

  Contains HM2007 (ANN voice/pattern recognition chip), HM6264 (64k SRAM, the Blonde Brain),

  a keypad for user interaction (train data, clear data, address select), a PIC16F84, another power supply and some jumpers.

  These pages are scanned from my old magic spell notebook. It was drawn many many years ago, at the dawn of internet era in this part of the world (at the end of the 90s).

  So for this project I'm using an old board I designed and made during the old days of CadSoft Eagle 4, when I did not know that program could handle data buses.
  

  Schematics look better on those papers.

  So how does it work?

  Let's say we are trying to recognize some voice commands. At system startup, there will be two leds active: a red led (pwrOK) and a blue led (READY).

  First we need to train the chip so we press the address (for example 01). At this time the RDY line go HIGH (Blue LED off). Pressing "TRN" will turn the RDY line back to LOW (Blue LED on) - ANN is waiting for an input - and we can speak the word. WLEN (word length) jumper is set to 0.9 seconds, so according to the datasheet we've got 40 locations available in the SRAM memory to store the training data [spoken words].

  When training the ANN with data (patterns, or words, or voice commands), we need A LOT of training data. Because if we train it saying "three", it will react positive when we say "thirteen" or "thirty". If there's not enough data, it cannot make the differences.

  Now - working with repetitive patterns (vibrations) is easier for the ANN. But if we are not careful, we get the "word too long" error, so this means we have to periodically separate the ANN input line from the vibration source. This is where the relay input select board is coming.
  

  Alright. Now what's up with that PIC16F84?

  Well, it performs two functions:

  - 1. getting the address bus numbers and do something with them - not implemented yet;

  - 2. When RDY signal gets low (this usually means keypad interaction), the 0.9s/0.3s...

  Read more »

• Science Time (IV): Dacia 1310 Engine Exhaust Gas Analysis using TCD Gas Chromatography

  [skaarj] • 06/20/2016 at 16:23 • 4 comments

  Good Morning and welcome to the 4th edition of [skaarj]'s Science Time log.

  Today we are going to review the analysis results using the rebuilt/repaired gas chromatographs described in the previous log. The experimental setup is described, and a set of results is discussed.

  From the beginning we have to set the things clear. At first I almost started to believe that my planned Citizen Scientist approach for this project failed for good. Somehow the Divine laughed at me and said something like "Let there [skaarj] hack his way out of this" and I was given the Chromatograph, and the rest of the story is known. The device is "calibrated" using expired gas calibration tanks, it was rebuilt in a hurry using questionable methods such as FID to TCD conversion plus electronic parts and wires found in the drawers. Rats and humidity tampered with most of the wiring and - most important - there is A LOT OF NOISE in the output signals and the precision of this monster is unknown.

  1. Experimental Setup
  

  For the beginning, this is the block schematic of the proposed Exhaust Gas Analysis system.

  This is my proposed block schematic for presenting a generic gasoline ignition engine and my way of tampering with it. Carburetor stays on top of the intake manifold and performs the air-gasoline mixture. Additionally there is a hydrogen-oxygen gas source which is mixed with the air.

  Input mixture goes into the engine, but there's a small hose (blue color in the figure) going to the first Chromatograph unit (Chroma In).

  A small part of the exhaust gases is extracted through a second hose (red color in the figure) going to the second Chromatograph unit (Chroma Out).
  

  Both of them perform the gas analysis as described at the beginning of the previous log.
  

  Output signals from both Chromatographs are connected to both the paper graphic recorder and a data acquisition card connected my old laptop to dump all the numbers.

  From left to right, top to bottom:

  1. Chroma In, Chroma Out, Paper Graphic Recorder (no paper), Spare Chroma;

  2. Hose in the left: air intake line; hose in the right going to the back of the car: exhaust line;

  3. Those two strange things near the right side container door: hydrogen electrolysis reactors;

  4. Dacia 1310MLS, made in 1986, engine capacity 1297cc, maximum speed (when wind blows in the rear) 100km/h; time to reach from 0km/h to 100km/h: around 10 minutes; engine pollution factor: SKY HIGH.

  Dacia 1310 engine. Still looks nice and still doing its job after so many servicing years.
  

  Big hose - hydrogen from first electrolysis reactor connected to air intake hose (black thing curved like a glove);

  Big transparent hose - hydrogen from second electrolysis reactor, connected directly to engine air intake manifold under the carburetor;

  The big black round cylinder (bottom side) is the air filter. In its left side there is a glass receipt containing the coolant (distilled water). Air intake hose sits on this cylinder, and the other end covers the carburetor air intake.

  That red paint on top of the air intake hose is liquid silicone covering a big hole.

  Small white hose near the hydrogen hose: Chroma In Gas Line.

  Two hoses inserted in the exhaust pipe. Those are going to Chroma Out.

  Welcome back to the Madhouse. This time we get to visit the yard.
  

  Hydrogen Electrolysis Reactors

  Construction and functionality of these electrolysis reactors is described in Science Time (II) log.

  They are electrically connected in parallel and powered at 220 volts direct current. Not AC, but DC, rectified through some HUUUUGE diodes used for welding. I also have 110V in the cabin but it's not so funny to use.

  Both electrolysis reactors are filled with distilled water and 1.5% NaOH. Unfortunately, Potassium Hydro Oxide (KOH) has a better efficiency in producing hydrogen through electrolysis but this product is strongly regulated in the European Union due to its... dangerous properties. This time there's no more police supervision so I must be careful and do all the things right.

  As...

  Read more »

• Science time (III) Giving second life for an ancient Gas Chromatograph

  [skaarj] • 06/09/2016 at 11:57 • 2 comments

  Greetings, fellow hackers.

  Now I (really?) understand the philosophy of the jury. Push everyone to hack more. If you have success with your project, push for more and rise up. If you fail, push for more and rise up. Never give up. Never let yourself go down. This is the true hacking philosophy: if you fail, or you are pushed away to give up, nobody cares. There will be other people ready to take your place and push further for nothing.

  Instead of receiving some bucks in order to get access to latest tech, now I must hack my way to the 3rd round prize by resurrecting some ancient tech dinosaur in order to get things done.

  So from this moment I am no longer questioning the second round results.

  Allright, Science time.

  I need to "smell" the exhaust gases from the engine in normal working conditions and in the presence of hydrogen-oxygen mixture injected through the air intake manifold.
  

  And for that I need a Chromatograph, so let's see what this device does and how it works.

  This is a block schematic of a generic Chromatograph:

  
  A light gas is used as a sample carrier - usually Hydrogen (flamable!) or Helium (preferred). The sample is injected before the column (some long metalic hose) which is supposed to be heated. Heat can also be applied to a second column (not shown in the figure). The column hose is extremely thin - with microns in diameter. Lighter gases come out first, heavier gases come out later. There is also some theory about gases being attracted or repelled by the column internal coating, which also influences how fast they are separated and how fast they travel through the column. The sampling time lasts for minutes, then the TCD (thermal conductivity detector) or FID (flame ionization detector) outputs some analog signal at a corresponding time, which is the percentage of compound (separated) gas coming out. The higher the spike recorded on the graph, the higher the concentration of that separate compound is. The later it is recorded, the heavier the gas is.

  The most common sensors (detectors) are as following:
  

  TCD - thermal conductivity detector. If some gas passes around a electrically heated tungsten filament, its temperature modifies (usualy increases) and its electrical resistance also increases. The carrier gas alone (helium) has high thermal conductivity and keeps the filament cool, maintaining an uniform electrical resistivity. When the carrier gas brings a foreign gas, resulting gas mixture thermal conductivity decreases so it does not allow all the heat to dissipate (as before) and forces the filament to get hotter. Its electrical resistivity increases and the corresponding current variation is recorded as a response.
  

  FID - flame ionization detector - it has two electrodes placed close to a flame fueled by carrier gas + air (in this case - hydrogen). This detector is used to identify the presence of hydrocarbons, so these are pyrolized by the flame. The hydrocarbons will split into ions (at cathode) and electrons (at anode) which generates a current between the elctrodes. It does not work with other gases (non-hydrocarbons) as these are unable to pyrolize.

  Both sensor responses are recorded as spikes on a graphing device, which form a Chromatogram of the sample gases analyzed.

  There are also catalytic sensors, infrared sensors, mass spectroscopy detectors and a lot more. All detect gases in their way but the output is recorded as spikes.

  That's it. Simple in theory, headache to build - or to pay for an expansive one.

  As I announced yesterday on #Hacker Channel, I found a 40-years old Chromatograph Station inside an abandoned warehouse at my work place. It's a true monster. They told me it worked with vacuum tubes and it's entirely trashed. During immemorial times, many generation of rats made it their home....

  I'm happy the camera cannot record the horrible smell.

  From top to bottom:

  - Hole Depth display;

  - Mud Temperature in, Mud Density in, Mud Temperature out, Mud Density out (there's a special prepared mud circulating...

  Read more »

View all 31 project logs