Shop OBEX P1 Docs P2 Docs Learn Events
Propeller P8X32A Radiation hardness? — Parallax Forums

Propeller P8X32A Radiation hardness?

Dr. MarioDr. Mario Posts: 331
edited 2011-07-23 10:57 in Propeller 1
After thinking about probably making myself my own storage (event-recording) geiger counter that's intelligent, and has organic LED display to show the radiation strength / counts per minutes, I am truly wondering how much Sievrets per seconds (since P8X32A runs much faster, I went with uSv/Sec - Sv/Sec as I need to emphasize on rapidly-changing events, which this geiger counter would be for, recording what's going on, onto the X-ray - proofed SD card, if the claims is to be believed, though) before P8X32A gets hit with Single-Event Upset (SEU), a transistor-level phenomenal which will cause the CPU to freeze or go loopy.

I am curious, if anybody have done radiation torture on P8X32A, I would be interested to know the results for numbers of hits, and how much milligrays was imposed on P8X32A before it go nuts.

Comments

  • Dr. MarioDr. Mario Posts: 331
    edited 2011-07-19 14:51
    It seems that there is no solid answer, so probably someday, I will have to torture a few P8X32A chip with the weak radiation source, Americium as found in smoke detector, and probably X-ray tube. After a while, I may have to perform the complete failure analysis, by hitting P8X32A very hard with FXR (I can build the flash X-ray generator, not too difficult for me - just too dangerous due to incredibly powerful capacitors rated at 10,000 volts or much higher and multi-kilowatts to few megawatts hard X-rays depending on tube size and voltage / joules of the capacitor banks) - to see how much hits the P8X32A can withstand, both in operation and powered off, before the chip latches up or worse, fries. The test here won't be too cheap, but at least it's for the good of the others...
  • Mark_TMark_T Posts: 1,981
    edited 2011-07-19 16:52
    I am reminded of "Will it blend?"...
  • Dr. MarioDr. Mario Posts: 331
    edited 2011-07-19 17:04
    That's kinda irony. But, good point, though.

    The reason I asked if the P8X32A would be able to withstand specific amount of Gamma-rays before it go crazy was because I may want to build my own storage geiger counter to keep track of emissive energy over time, like from few seconds to few days. Apparently, there is no solid answer as of now, so I kinda decided that I will have to find out on my own, and probably post the result for the others...
  • HumanoidoHumanoido Posts: 5,770
    edited 2011-07-19 17:07
    My project requires protection against x-rays, gamma rays and cosmic rays. Like any processor chip going into space that needs to be "space hardened," the chip can be lined with a piece of lead. I would recommend determining the radiation parameters such as depth penetration of the most energetic particle and then determining the required thickness of lead so as not to damage any propeller chips during the project.
  • Dr. MarioDr. Mario Posts: 331
    edited 2011-07-19 17:20
    I see. Yep, depleted Uranium can also be used as a shield (it's electron-heavy which will absorb even shortwave Gamma-rays, but is much heavier and much more expensive than normal Lead shields), even though I was a bit curious if the P8X32A can stand about 0.6 Grays before it starts to behave funny. I will install Lead shield inside the handheld body (a project box), though, as the sensitive electronics such as Op-Amps will need to be shielded, or else the false reading will really ruin your day.

    I have heard that some SD are designed to be Radiation-Hardened (as the airport X-rays machine really do great job at bulk-erasing some types of SD cards), I am not sure, but if it can be confirmed, alright then I will use it in this geiger counter.

    And for dirty method of testing the thickness of Lead for required energy absorption, you can use cheap MOSFET, such as IRF510, and then observe its behavior as you hit it with X-rays or Gamma-rays, if it ain't latching, then thickness is good enough, although I would tend to derate that by 10% to buy me some space in the risk assessment.
  • RaymanRayman Posts: 14,876
    edited 2011-07-19 17:54
    I think DU would be a poor shield because it's a bit radioactive... Maybe more than a bit...

    I think you could determine the rad-hardness of the Prop based on the "process" from which it was made.
    There are probably lots of chips made with the same process, all with about the same level of hardness.
    My guess is that this process is not specially rad hard.
    There are SOI processes that are more rad hard used for space applications...
  • ElectricAyeElectricAye Posts: 4,561
    edited 2011-07-20 06:54
    Mario,

    I asked about something similar some time ago, so you might look at this;

    http://forums.parallax.com/showthread.php?130448-Is-a-DIY-radiation-detector-possible-using-memory-devices-such-as-EEPROMs-SD-cards&p=985655&viewfull=1#post985655

    Especially the comments of localroger might throw some light on this.
  • Pharseid380Pharseid380 Posts: 26
    edited 2011-07-20 09:49
    Humanoido, have you checked out AMSAT? They're HAM's who've made their own satellites. I would imagine they're a pretty good source for information in this area.

    -phar
  • Dr. MarioDr. Mario Posts: 331
    edited 2011-07-20 10:02
    ElectricAye, THANKS! It got useful information, and now I am convinced that my P8X32A chip (current generation of P8X32A die - 180nm SOI) can withstand even fatal dosage of radiation. Good enough for my geiger counter.

    Rayman, it may be true, depending on purity of elemental Uranium. However, as mentioned here, http://en.wikipedia.org/wiki/Depleted_uranium military and medical equipments have been using it for a while to curb all of those nasty radiation, possibly several hundreds times the LD50 fatal dosage of human model, or way higher. I will still go with Lead as it's much cheaper and lighter (and also more malleable - DU will have to be melted and poured into ceramic mold).
  • hinvhinv Posts: 1,255
    edited 2011-07-20 11:38
    Yeah, depleted uranium is a misnomer and probably intentional. Otherwise, how wuold you get unsispecting sildiers to destroy their ability to have normal children.
    Do a google image search for "DU babies" to see what I mean.

    I thimk your project is a worthwile one. I think you will have to do constant memory testing of hub ram and cog ram. Does anyone know if there is internal parity memory in the prop? I haven't heard of it mentioned, so I doubt it.

    Doug
  • Dr. MarioDr. Mario Posts: 331
    edited 2011-07-20 18:33
    hinv, Yep. DU is also classified as carcinogenic - more aggressive than Lead, due to its unstable chemical characters (some chemicals in your body would be looking for some more free electrons, and DU got plentiful - a fatal consequence), so your comments is a valid one. The purchasers of molded DU are responsible for giving it biological protection, such as after assembly, they will need to go in plastic (or metal) housing of the equipments being shielded. The DU dusts will need to be sucked up by a special vacuum cleaner, such as the one used for printer toner, only fitted with higher-quality HEPA filter - the assembled hardware MUST be spotless afterward.

    Therefore, DU is only used as a last resort if thick Lead shield won't work - CT X-ray tube don't use this kind of material, they're packaged in thick Lead alloy material - DU also heat up much when it starts to absorb 800W X-ray energy, and it's a crappy heatsink material, so Lead + Copper (probably with trace of Tungsten) alloy is used instead for this kind of X-ray equipment.

    Also concerning battery type as mentioned here http://forums.parallax.com/showthread.php?132961-Propeller-Power-Consumed-Based-on-Code , my geiger counter would be powered off by a Lithium-ion battery so it would last my measurement long enough (for safety reason, what's good of a geiger counter with expired Alkaline batteries) - P8X32A consumes lesser power so it's a nice asset.

    How will I have it detect radiation? There are two ways to do it: Geiger-Muller tube and / or PIN photodiode. Geiger-Muller tube, either sealed or axial flowing type, detects radiation (X-rays, Gamma-rays, Alpha [Proton], and Beta [Electron]) by ionizing noble gas (Neon, Argon or Xenon, depending on type of counter) distilled in Iodine vapor, and once radiation hits the noble gas atoms, they start to conduct electricity upon the path of radiation and would be quenched by Iodine, bringing it back to non-conducting state to be repeated - the way Geiger Counter works.

    PIN Photodiode works similarly to GM tubes, only they are safer to touch than the tubes (in most case, though. Some go up to 200V) - they're shielded by thin Aluminum foil (like the one used to be wrapped around the food) to keep long-wavelength lights (Microwave to Extreme UV) out to prevent false reading, allowing only X-rays and Gamma-rays to pass through and latch the photodiode. One such example - http://www.maxim-ic.com/app-notes/index.mvp/id/2236 - as mentioned in the application note, you will have better chance if the photodiode has large light-sensitive area (more larger PIN light-absorber the better chance you will have at detecting very short light wavelength - and it will work better if it's packaged in plastic package.) An example of plastic-packaged large-area PIN photodiode - http://www.mouser.com/ProductDetail/Vishay-Semiconductors/BPW34/?qs=sGAEpiMZZMtWNtIk7yMEsVnpcClC5AWfAIjz0gFScyY%3d - I know it's Mouser, but it was just an example. And, wavelength sensitivity is not going to matter much anymore after passing 20nm, as it got plentiful of energy, thus having higher chance of knocking off two-electron pairs. Your digital camera(s) are also sensitive to X-rays (in certain brightness - a very powerful X-ray tube, such as the one used in NDT equipment will cause it to snow much - blinding it completely).

    I would probably go with solid-state option - Geiger Muller tube may need to be included too, just in case, and I may need to ground Lead shielding as it would be one outrageously sensitive RF receiver, not what I would want seeing on the OLED screen - a very bad thing. They also will need to be calibrated by firmware after booting up. SMPS will do fine here as long as you do generous filtering on its resultant DC rail and Aluminum Oxide polymer capacitor would do great job at acting like a battery upon SMPS' off-state.
  • Dr. MarioDr. Mario Posts: 331
    edited 2011-07-21 14:03
    I may use the yellow OLED screen; http://www.mouser.com/ProductDetail/Newhaven-Display/NHD-27-12864UCY3/?qs=sGAEpiMZZMshRHSPqbouvK1hwgkEjgCbtjNxgg21nz8%3d as it's tad easier on eyes (blue's tad too sharp for some peoples - and night vision preservation is also very important so...) to display the data. If you guys think it's better to go with Vacuum Fluorescent Display, I can change my design layout.

    Also, you guys think it's good idea to put in SMD fuse holder holding a 1.25A Instant-blow fuse (http://www.mouser.com/ProductDetail/Schurter/3422001011/?qs=YybYKR%2fZNJEk9Jbe8NJHig%3d%3d) in line to Lithium-ion battery's anode connection to the intelligent Li-Ion protector chip, just in case? I thought it would probably be a good idea to fuse the battery anyways, as the total failure of Li-ion battery protector do happen, although very rarely.
  • Dr. MarioDr. Mario Posts: 331
    edited 2011-07-21 21:04
    Here, I doubt parity could be done for P8X32A's on-die RAMs (COGRAM / Hub RAM), but it's possible to do quick CRC checksum on small pages, and it can be discarded if it's screwed up due to neutron impacts (which flips the RAM transistor state from 0 to 1), and I will just shield the P8X32A with grounded Lead shielding (even when grounded, it could possibly repeal Beta particles and absorb Alpha particles - Lead is already good for it, but grounding may improve it anyhow) - and for some chips, like programmable Op-Amp chips, I probably may have to do dead-bug technique to prevent false reading (and it makes probing the SOIC chips easier anyhow).

    And, of course, generous filtering on SMPS will have to be required (any switching noise is unacceptable, but I need to make a gamble here, though - battery life is very important.... And I dunno if most ICs will be happy with less than 3.3 Volts DC power...) and many digital chips or even RF-class chips will need to have low-ESR 100nF MLCC [multi-layer Ceramic capacitors] to be able to keep the RF-induced current from causing some weird issue.

    I would also hook up a pin from P8X32A to the output of amplifier to count the "events", that it counts the time the heavily biased photodiode or geiger-muller tube, thus allowing the P8X32A to truly count it out and print the events onto the display or write the timestamp events onto the SD card the moment it "collect" the values. It also simplify the counter calibration, the moment it's turned on (compared with stored calibration reference values on the firmware ROM).
  • Dr. MarioDr. Mario Posts: 331
    edited 2011-07-23 10:57
    I think I probably may have to put this thread on hold as there are not much interest anymore, so I would just leave it at that until I get the geiger counter done, and post the results.
Sign In or Register to comment.