Advertisement

Blog

Can We Have Rad-Hard Integrated Analog?

Two recent blogs about radiation and analog integration have me wondering about the practicality of large-scale integration for space applications. The first blog was written by Paolo Scalisi. He discussed the mechanism by which radiation can cause ICs to temporarily or permanently malfunction. This blog was in three parts. The third part is here; you can get at all three parts from that link.

The second blog was written by Steve Taranovich. He discusses the need to squeeze lots of functionality into a small space for applications related to space travel. Obviously, you want to keep the weight of your launch vehicle as low as possible, so small ICs and small PC boards are needed. Power management ICs and analog front ends (AFEs) are a couple areas where Steve suggests we should concentrate our efforts:

First of all, it takes larger rocket engines and more fuel as onboard weight increases on a mission, so you want to keep weight and size down to keep costs down. For an electronics payload, guidance, communications, and control systems that means high levels of integration… AFEs, Power Management Integrated Circuits (PMICs), and microcontrollers with lots of analog integrated or lots of analog with a micro integrated are recommended. You also want efficient power — avoid heavy heatsinks and transformers if possible. This means switching regulators, low resistance MOSFETs…

My concern with regard to putting all this functionality on a small piece of silicon is that you are obliged to use some very fine line processes and make devices that have very low breakdown voltage ratings. This means the problems to which Paolo referred in his blog will be exacerbated:

To understand the effect on an IC let’s examine the effect of the radiation on a power MOSFET switch, which can be considered an elementary component of an IC. Due to ionizing radiations, e /p+ pairs are created inside the gate oxide (SiO2 ) and the electrical field moves the more mobile charges to the gate terminal. Meanwhile, the holes (shown as h+ in the Figure) remain inside the oxide. Some of the electrons that flow from the source to the drain will recombine with the holes inside the gate oxide, resulting in some holes disappearing from the oxide.

Paolo goes on to explain how this can cause a shift in the gate threshold voltage of the MOSFET that had been irradiated. At least the FET would still function (though perhaps not quite properly). Beyond merely a threshold shift, you could expect the ultra-thin oxide layers would be damaged.

My conclusion here is that we may be stuck with the older processes for fabricating devices intended for high-radiation environments. What do you think? Can we do large-scale analog integration of rad-hard parts?

— Brad Albing, Editor-in-Chief, Planet Analog and Integration Nation Circle me on Google+

Related posts:

49 comments on “Can We Have Rad-Hard Integrated Analog?

  1. SunitaT
    December 31, 2013

    My concern with regard to putting all this functionality on a small piece of silicon is that you are obliged to use some very fine line processes and make devices that have very low breakdown voltage ratings.

    @Brad, I think its better to avoid doing large-scale analog integration of rad-hard parts because putting all functionality on small piece of silicon can also cause lot of cross talk noise.

  2. RedDerek
    January 2, 2014

    BradA – I recall that the smaller the feature size, the more susceptible to radiation. Thus most space electronics would use a large feature size or planar technology. I also post on the part 2 of Paolo's articles, I believe most of the space based electronics are using the large feature size processors and stuff.

  3. Brad Albing
    January 2, 2014

    @RedDerek – thanks. And for anyone wanting to see that line of comments to which RedDerek referred, go here:

    What Do You Need for High-Rel & Rad-Hard Applications? Part 2 – comments

  4. etnapowers
    January 3, 2014

    @Sunita: the crosstalk is a little price to pay in term of cost saving due to large scale integration. The price of a power switch inserted in a rad hard package is more than 400 $

  5. etnapowers
    January 3, 2014

    The design strategy of rad hard device for a large scale integrated IC should be based on the protection of those parts of the devices that are more sensitive to the radiation, this is not so easy but it's the right direction to be followed.

  6. Davidled
    January 5, 2014

    If cross talk would be a concern, the data rate for TX and RX could be reviewed because cross talk could be generated in the high speed transmission above 5 GHz. Noise is slightly different from cross talk.

  7. etnapowers
    January 8, 2014

    @DaeJ: that's correct , normally the transmission rate has to be high so the crosstalk may arise, the layout of the aerospace module is really important to avoid the crosstalk between TX and RX and to make the communication effective.

  8. etnapowers
    January 8, 2014

    The system engineer role is very important to solve the crosstalk issue. The system layout and the PCB design strategy ( path insulation by ground planes for example) can minimize the crosstalk, so the module functionality might not be affected in case of a proper engineering process.

  9. fasmicro
    January 8, 2014

    >> putting all functionality on small piece of silicon can also cause lot of cross talk noise.

    If you are good designer, you can control that in silicon. That is not the main problem with rad-hard design. When a chip has to work in the space environment, you need a new process and a new design paradigm. Fundamentally, the way your transistors work will change and you need to influence that at process and new design approach. The normal cross-talk is not the big headache here. The key is that your chip will just die when put in these states if not engineered for the environments.

  10. fasmicro
    January 8, 2014

    >>  I recall that the smaller the feature size, the more susceptible to radiatio

    This is a solid state physics which means at the end all the transistors will be integrated. I do not understand the fundamental physical reason why the larger feature size works better for space electronics. Any help here?

  11. fasmicro
    January 8, 2014

    >> the crosstalk is a little price to pay in term of cost saving due to large scale integration.

    The key contributor of cross-talk may not be largely due to the level of integration, rather on the feature size of the transistor. If you moved down the nanometer size CMOS, you will have more cross talk even if the integration is not very large. Conversely, if the integration is very large and you use large feature sizes, you can reduce the noise. Capacitive noise, i.e. the parasitics, are more dominant in the nanometer regime and drives the noise budget.

  12. etnapowers
    January 9, 2014

    @fasmicro: I absolutely agree with you, the good designer can avoid the cross talk by utilizing a good design strategy , the real challenge is to protect the device from radiation that can cause a failure, hence the designer will have to work closely with the product engineer to be sure that the device is radiation hard qualified.

  13. etnapowers
    January 9, 2014

    @fasmicro: the larger feature size works better for space electronics because the damages introduced locally in the silicon structure may not reach the active areas of the device and this results in a lower failure rate.

  14. etnapowers
    January 9, 2014

    @fasmicro: nice post, I think that the crosstalk can be minimized by a good compromise between integration and feature size, the crosstalk is not the only source of noise expecially in a radiation hard environment hence a good layout design of the application board and a proper protection to the radiation, will ensure a good working of the aerospace module silicon based.

  15. etnapowers
    January 9, 2014

    The SoC can be protected to the radiation by mean of a radiation hard macro-package having a frame resistant to radiation and a good design strategy of each component of the SoC, that has to be designed in order to protect the parts of the device that are more sensitive to the radiation effect.

  16. Rajan Bedi
    January 10, 2014

    The space electronics industry is already exploiting the benefits of rad-hard, integrated analog, digital and mixed-signal integrated circuits. The following state-of-the-art payload processor uses a number of strategic integrated technologies fabricated using CMOS, bipolar, SiGe BiCMOS and GaAs processes:

    http://www.edn.com/electronics-blogs/out-of-this-world-design/4419770/Alphasat-s-Champagne-moment

    For those of you who have an interest in space electronics, I have a monthly blog at EDN called, Out-Of-This World Design .

    http://www.edn.com/electronics-blogs/4406636/Out-of-this-World-Design

    Each additional kg of mass adds $50k to the launch so our industry has to exploit the power, size, mass and reliability advantages of integrating technologies to deliver microelectronics that will function in GEO for up to fiftenn years. Once operational in orbit, we will never be able to physically handle our electronics again. Every mW of power counts!

    If you wish to be added to the mailing list of my monthly blog on space electronics, send an email to:

    Rajan.

     

  17. goafrit2
    January 13, 2014

    >> hence the designer will have to work closely with the product engineer to be sure that the device is radiation hard qualified.

    That is good point. One of the reasons why not many people work in this area is that everything is specialized. From the transistor to the process, you cannot use the traditional CMOS technology. Apparently, they have to command premium when they make these products for the aerospace industry.

  18. goafrit2
    January 13, 2014

    >>  good layout design of the application board and a proper protection to the radiation, will ensure a good working of the aerospace module silicon based.

    Always be careful. I was told if you work in this area and come up with some really breakthrough performance metrics in your design, you may lose the ability to patent it and government will rather pay you off. This will be by far the most lucrative area in the industry for those that can invest in the techology to become leaders in it.

     

  19. etnapowers
    February 3, 2014

    From your link:

    “The spacecraft was developed as a result of a public-private partnership between ESA and Inmarsat and will live at 25°E in geostationary orbit covering Europe, Africa and the Middle East. The satellite weighed 6,605 kg at launch, has a size of 7.1 m x 2.5 m x 2.8 m, consumes 12 kW”

     
    @Rajan, it's been a great experience for you, true? What kind of reliability issues did you encounter for the development of Alphasat satellite?
    What electronic blocks were more exposed to radiation damages?
  20. etnapowers
    February 3, 2014

     I work with ESA to develop rad hard power switches to be inserted in power management modules on aerospace applications. I think that it's a really promising sector for electronic industry whose interest in this type of application is fast growing

  21. etnapowers
    February 18, 2014

    @goatfrit2: I fully agree with you, the traditional CMOS technology is not enough robust to deal with radiation , so you have to use a specific technology for the silicon and morover a specific technology for the packaging step.

  22. fasmicro
    February 19, 2014

    >>  What kind of reliability issues did you encounter for the development of Alphasat satellite?

    Certainly he cannot provide such answers here, realistically. That is more than a PhD thesis. Nevertheless, it is obvious that radiation will be an issue.

  23. fasmicro
    February 19, 2014

    >> . I think that it's a really promising sector for electronic industry whose interest in this type of application is fast growing

    You need to be prepared for the regulatory challenges in this sector before you dive to take advantage of the perceived opportunities. It is like the automotive sector where everyone sees opportunities but regulation kills most of the startups that attempt.

  24. fasmicro
    February 19, 2014

    >> so you have to use a specific technology for the silicon and morover a specific technology for the packaging step.

    One of the reasons universities lag behind in this area of research. People need a lot of capital investment in new process, technology etc to build rad-hard systems. And the profit yield could be great if one gets it right. It is like the medical electronics that has great margins for the few that can do it. But medical electronics is luckier as the technology does not need to be wholly new as in rad-hard.

  25. etnapowers
    February 20, 2014

    @fasmicro, I was asking only a short description of some of the issues encountered, I agree with you on radiation issues.  I would like to know what kind of SEE (Single Event Error) did appear on the electronic circuitry: for example SEGR (Single Event Gate Rupture) , or SEU  (Single Event Upset) or what else?

  26. fasmicro
    March 12, 2014

    >>   I would like to know what kind of SEE (Single Event Error) did appear on the electronic circuitry: for example SEGR (Single Event Gate Rupture) , or SEU  (Single Event Upset) or what else?

    Most are classified. The bad deal is that some of the great works in this area cannot be patented because government will not allow you to share that information to the world.

  27. etnapowers
    March 13, 2014

    @fasmicro: I have recently written a blog series here on the Planet analog, that is on going to be published, here you can find the first part, I hope this could be interesting for you.

  28. etnapowers
    April 1, 2014

    @fasmicro: the regulatory challenges are really tough in this field but this is due to two main reasons:

     
    – the level of reliability that has to be guaranteed
    – the hard environment in which the ICs have to operate
     
    The gain margin is however very good so I think that it is a promising business.
  29. etnapowers
    April 1, 2014

    @fasmicro: you're right, it is very expensive to invest in the rad hard technology, but I think that if a company creates a good portfolio of products for aerospace applications,  it has the chance to become a leader in a really promising and high margin business. Just bet and win.

  30. Netcrawl
    April 1, 2014

    @etnapowers you're its a very expensive business. The aerospace and defense industries demand an ever increasing functionality and capabilities in space-based components and equipments such as satellites and space-deliver vehicles.

    @etnapowers What about RHBD? RHBD suppliers (Rad-Hard by Design) promised space-ready, radiation-hardened space components, ending a decade long challenges in aerospace industry. RHBD is a viable design approach for producing space application integrated circuits, several companies have adopted this new business model in the past 2-5 years to bring advanced integrated circuits to the marketplace and meet the requirements of the satellite marketplace.   

  31. etnapowers
    April 2, 2014

    RHBD is a really promising solution , provided an effective design based on the test results coming from the radiation facilities. The new design companies will be successful in case of patents released after the irradiation tests performed in the facilities.

  32. fasmicro
    April 13, 2014

    Thanks for the pointer – very appreciated. Nice piece.

  33. fasmicro
    April 13, 2014

    – the level of reliability that has to be guaranteed

    – the hard environment in which the ICs have to operate
     
    The gain margin is however very good so I think that it is a promising business.
     
    Generally, that is how the industry works and I agree. When you make a phone, if the gyro fails, you can throw it away. That is not the case when you use a gyro in a car. There is a danger that lives can be lost.
  34. fasmicro
    April 13, 2014

    Just curious, do you know of any speciality IC firm that focuses on this rad-hard subsector in our industry. I know that TI, ADI and co may have pockets of units for this; yet, I want to know if we have firm-level specialties.

  35. fasmicro
    April 13, 2014

    . RHBD is a viable design approach for producing space application integrated circuits,

    I have made many RHBD circuits for floating gates (analog memory). The major problem is that even after the special circuit architecture, you still need a special process for it. The circuit does not solve the problem. In other words, even a RHBD in a typcial CMOS platform will not qualify for a space product line unless the process is as they want.

  36. etnapowers
    May 6, 2014

    @fasmicro: you're welcome, I provide you another link , hoping you appreciate it.

  37. goafrit2
    June 4, 2014

    The complexity of the electronics industry with the drive to reduce cost has made it difficult for new entrants to enter the game. The reality is that electronics startups cannot simply use prototypes to win contracts unlike for web and app counterparts. Experience and history matter in electronics and that is why we have legacy factor in the industry. It makes life easier for the established players since disruption cannot be too sudden.

  38. fasmicro
    June 7, 2014

    Thanks. I have read that article on aerospace failures. These are special sectors with specific domain expertise. This makes it exciting when I read what SpaceX is doing in the industry by evolving as a competitor within years. Very hard stuffs indeed.

  39. fasmicro
    June 7, 2014

    And when you move into the highly regulated aerospace industry, you will appreciate what SpaceX has accomplished by crashing the parties. You need to have so much quality for investors to believe in you in some of these sectors since technology is not just the solution, regulation is a key part.

  40. Davidled
    June 7, 2014

    Unless the company has a reputation in the market, three factors would be required:quality;technology;and cost. In the most market, I think that clients are looking at these three criteria. It could be a very tough for startup company to appeal to customer unless company meet these three components. Regulation would be related to all three.

  41. goafrit2
    June 9, 2014

    >> It could be a very tough for startup company to appeal to customer unless company meet these three components

    Yes, and add regulation. SpaceX is seeing that you can still meet the three on rockets and satellite launching and still be out of luck with DoD because of regulation.

  42. etnapowers
    June 13, 2014

    @fasmicro, you're right, it's a very competitive market and it's very difficult for a company to enter into this business, because it's very hard to find some experience on the radiation effects on ICs. I guess that many years of trials and studies are required in a demanding business like the aerospace one.

  43. fasmicro
    July 3, 2014

    >>  I guess that many years of trials and studies are required in a demanding business like the aerospace one.

    The only exception is SpaceX which crashed the aerospace party within years.

  44. etnapowers
    July 7, 2014

    @fasmicro, it's really interesting the link to SpaceX:

     

    “SpaceX has gained worldwide attention for a series of historic milestones. It is the only private company ever to return a spacecraft from low-Earth orbit, which it first accomplished in December 2010. The company made history again in May 2012 when its Dragon spacecraft attached to the International Space Station, exchanged cargo payloads, and returned safely to Earth — a technically challenging feat previously accomplished only by governments. Since then Dragon has delivered cargo to and from the space station multiple times, providing regular cargo resupply missions for NASA.”

     

    This company collaborates with NASA that has the IPs for aerospace IC products for aerospace modules, it's a good example of application of the Research to the aerospace industry.

     

     

  45. fasmicro
    July 9, 2014

    . It could be a very tough for startup company to appeal to customer unless company meet these three components.

    That is the #1 barrier of entry in the hardware business. To be a hardware supplier, it does not always happen in the college dorm. It demands years of experience in a great company before the big buyer will give you a chance. That explains why it is not always common to see these firms come from college to great success unlike software.

  46. fasmicro
    July 9, 2014

    This company collaborates with NASA that has the IPs for aerospace IC products for aerospace modules, 

    Yet, SpaceX gets credit. NASA may have its own IPs, I think SpaceX has its own. You can attest to this by looking at the cost it takes SpaceX to build its system. NASA needs millions, SpaceX not so.

  47. etnapowers
    August 4, 2014

    I think that the collaboration between NASA and SpaceX is really profitable for both of the companies. I like both approaches, NASA has a scientific approach to qualify a portfolio of aerospace ICs, SpaceX is oriented to the market of spacecraft and so it's a commercial approach.

  48. fasmicro
    August 5, 2014

    I think that the collaboration between NASA and SpaceX is really profitable for both of the companies.

    Maybe you wanted to write “beneficial” since NASA does not report any profit.

  49. etnapowers
    August 6, 2014

    @fasmicro: NASA is  funded by the USA government, right? The benefits of NASA experience are enjoyed by people, for example one of the benefits is the study of the comets's trajectories and the prevention of interference with the orbit of planet earth, or the study of earth climate from the space perspective.

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.