Automotive Radar Millimeter-Wave Technology from Frescale

Automotive Radar Millimeter-Wave Technology from Frescale


well good morning welcome to this first presentation in this session my colleague Jim shocky and I will be talking with the next hour about the Ottoman motivator and the effort that we have at free scale for this application in terms of the technology development in terms of the components that will be required for that I will cover the front end the RF components for the automotive radar applications and gene we’ll be talking more about the data processing and streaming okay let me introduce myself I am dragon Medina which and I’m the operations manager from the sensors division at free scale I’m in charge of developing the front end RF components for the automotive Raider in my part of the presentation I will be covering mostly the technology that we are employing for this product I will discuss briefly the architectures that we are addressing or proposing for this application shortly along the same lines I will show some building blocks and the structures of our components and of course I’m not going to take much time showing the current status and results of the samples that we already have however you’re very welcome to talk to me after the presentation I will be around and I can definitely show you much more during this front and where we stand what they are planning to do so let’s have this now just as a preliminary introduction to what we currently have regarding the technology Freescale has been known as the technology company historically we were always in the front of the technology development and that stands now as well when we talked about the automotive radar applications what is happening is actually we observed that the gallium arsenide which is currently used in in the radar applications probably will go through the significant changes transition to the silicon germanium which looks like that as the most promising material for the next generation of these components and that’s exactly what we are trying to do however we are not developing a brand new technology the silicon germanium process flow to address these requirements but rather we are using the existing process flow existing process blocks and adding the new module in that process flow which would enable us to have the silicon germanium selective epitaxial stages that we can build only those devices that require to have this hetero Junction in order to achieve the high frequency 77 gigahertz capability with that said you can see here that we are using the 180 nanometer technology node originally it was the CMOS process to that basic process we added modules for various analog components and devices and finally we added the module which gives us the Selective epitaxial layer for our heterojunction what is really important here is that we have the bi CMOS process in a way that differentiates us from some other companies offering also silicon germanium but it is mostly the bipolar process so the differentiation that I am underlining here with the bi CMOS is significant because when we talk about transitioning from the gallium arsenide to the silicon germanium that is one level of the advantages that the silicon actually brings in terms of the pricing in terms of the processing the volumes however the other main point is the integration level and the integration of our bi CMOS process is exactly the advantage that we see with this process versus the just just bipolar process namely when I say bi CMOS it means we can build not just the amplifiers or the standard analog blocks but we can also add a significant digital domain and make these devices a high level integration devices very capable with additional features and functions and I will mention that a little bit later right now I am just saying we do have this technology we have been using it because it has been developed some years ago and this is exactly what we think brings us a significant advantage in the market space some of the cross sections of our heterojunction devices are shown here basically the history of the development obviously our first demonstrated type frequency devices were presented like ten years ago first time at the time it was the the capability of up to F max 110 gigahertz which as the time went by the enhanced improved and as you can see now on the right hand side our current heterojunction devices are capable of the F max almost at 300 gigahertz this definitely helps us in delivering good low noise performance for the 77 gigahertz operation some of the characteristics and measurements from this current technology and devices you can see characteristics are showing the F Max and F T as well as some correlation from the reported published data from some other manufacturers and universities you can see that Freescale with the peak F max actually correlates very well and is in the top performance of these type of characteristics let’s talk a little bit now about the architectures that we are trying to address here with this technology obviously speaking generically what is required or available out there when we look at the different regions globally we see various requirements and it’s probably expected because there are high class cars that there are mid class cars different regions target to different customers therefore the requirements for the complexity of these systems also vary and we can see that with these architectures as well the very simple mono static radar shown here obviously integrates as much as possible into a single device and even combines or reuses a single antenna for both transmitting out and receiving the signal in terms of the cost in terms of the price this is probably the least expensive architecture at the same time as some of the performance characteristics obviously are modest as well because of this combination and the cross docks and the noise level and so forth therefore for the higher end of these systems there is the so-called by static radar where the antennas for the transmission and receiving signals are separate and there is a much better insulation between the signal paths therefore the quality of the signal data processing distinguishing the targets or any other causes of the noise is much better controlled therefore the this systems type of systems here are much better performance and are also in demand by some of the customers so talking about these type of systems and reiterating what I said at the beginning regarding our technology capability to integrate both analog Digital Domain into a mixed signal monolithic devices what we are targeting here you can see here for the transmitter we have the transmitter which also incorporates the PLL in that monolithic signal die so the whole control for the VCO is within the device because of the digital domain the PLL we also have the state machine there which brings some sophistication or bring some features that can be programmed in and the device can be almost standalone functioning with the controlled ramp app ramp down features and so forth basically it removes some of the load that the microcontrollers used to have taking care of all the frequencies and ramping up and down therefore this device becomes very powerful with this type of integration and it can be customized it can offer in advance having different those functions therefore easily just through the spy you can request which which type of ramp up and down feature or mode you want and after sending that command once to the transmitter it will take care of it and run it in that mode of course whenever you want you send another spy command and you switch the modes of operation so that’s on the transmitter and I believe we will be the very first one in the industry now having this type of integration the PLL Ridgid Smita that has also the MPA – divider and some other sensors let me mention a little bit the receiver side what we are targeting obviously is multi-channel receivers and talking about complexity of these type of systems and raiders we hear different requirements the number of channels can go from 2 up to a dozen of channels either way we are capable of doing so the next generation of the receivers in addition to the multi-channel capability will bring also the integration of the ADC the converter so between the transmitter with the integrated PLL as the single die the multi-channel receiver as a single die these two components actually to die if you want basically represent the whole front end for the automotive raiders this block diagram goes back to the transmitter just emphasizing a little bit more the next level of granularity what are the main building blocks for the transmitter I’m sorry this is the receiver basically the error signal which is the local oscillator signal we are using between the transmitter and the receiver is the 38 gig which is probably easier to handle than the 77 on the board level so you can see the doubler and the mixer after that this is the simplest version of the receiver obviously doesn’t show the multi channels I’m not showing here the transmitter let me describe it a little bit the transmitter has other than what I mentioned that the PLL the state machine the VCO has NPA the the power amplifier has also the temperature sensor which is used for the temperature compensation because we are targeting automotive applications and that means the full capability from minus 40 degrees up to 125 Celsius ambient temperature right and finally I want to mention the power control the output power block which gives you actually a signal where you can control if over the lifetime out in the field even this device maintains its output power so that so that you know that over time it doesn’t deteriorate doesn’t doesn’t lose the power and the signal or the targets that you are measuring is really achieving that the distances that you are expecting so this is available as well as the divider that gives you also in the application a feature where you can make sure that the frequencies that you’re using seventy six and a half gigahertz or 77 after being divided so many times 1,000 times or so we are talking about the signals in the megahertz range so you can control those signals as well making sure that who are in the exact frequency range that you’re expecting this device to operate in any case there are many more details that I don’t want to dive into I’m available though outside after this presentation if you would like to discuss that more and I think I will turn it over now to my colleague Jean to talk about the data signal processing well thank you Dragan prescale has been in the wireless business for quite a long time and they were involved with a lot of the early development of the cig EC type of technologies that are being used in these radars that’s really where the exciting stuff is happening right now and we’re bringing a lot of that innovative capabilities that we’ve utilized in our wireless division for a long time to this new area and really the ciggy radar is positioning the automotive industry for a significant step forward and the ability to put these sophisticated systems into your cars and make it cost-effective for regular people like us to buy or probably some of you are here are driving BMWs already anyway and are not so worried about that cost aspect but for a lot of us the reality of these advanced systems is about to happen largely because of these technologies that were discussing so my part of this presentation has really focused more on the digital side now we’ve gotten the analog signal into the system what do you do with that how do you process it for those of you who might have come to the session last year on the radar it’s very similar coverage kind of limited to the specific signal processing that’s involved so some of this may be familiar to you but a quick overview of the signal processing that’s done in the radar system this is in the modulation techniques that are used by the transmitters and receivers of the of the radar pulse modulated oceans are really the original approach to radar and they are indeed used in automotive applications today mostly for short-range detection pre-crash type applications blind spot detection they’re very simple to implement simple electronics involved with generating the pulse in wave form but it’s also limited in its ability to detect range so it makes it good for short distance type functions now there are some approaches that are being used there was a public paper published just last year talking about how to extend pulse radar into more sophisticated applications like adaptive cruise control so there are some thoughts about how this can continue on and be a part of some of these more advanced systems in the future FM CW frequency modulated continuous wave modulation techniques are pretty much the more common approach to addressing the advanced systems like adaptive cruise control or pre-crash certainly any distance tracking that’s involved for an automotive situation like detection of an object out to 100 or even up to 200 meters distant it have fast enough rate to be able to start to modify the car’s behavior or track the objects around the car so that you can take effective action in relation to what they’re doing it has much better range accuracy because you don’t get that ambiguity that’s caused by the pulse transmission you can be continuously transmitting as opposed to a a pulse form of modulation where you have to transmit and then wait for the receive form FM C W your transmitter is constantly broadcasting it’s shifting its frequency though is a broadcast so you’re actually looking at the resulting information shifted in time somewhat so you don’t have the collision between the transmitted data or not now there are also techniques using frequency shift keying not so common in the automotive industry that I’ve seen although there are people who are talking about ways of combining some of the features of FM and pulse technologies to try to get some of the benefits of both some of the simpler electronics involved with the pulse type modulation so I’m showing right now is what you would see in an FM CW modulation scheme again this is the one that’s the most commonly used today so this is what the focus of the rest of this discussion is going to be when you’re talking about just detecting distance you’re really looking at just recognizing the transmission time of the pulse to its received time very simple and divided by two and you essentially have the distance to the object that’s reflecting the information so what you see in these diagrams and we’ll start with the far left your far left is this the simple frequency modulated waveform this would be in our case around 77 gigahertz typically the the bandwidth that’s used to transmit the the frequency modulation component is determined by the amount of the distance you want to to track but typically for our applications that’s around you know megahertz that we’re interested in looking at the distance on so the blue is the transmitted signal and the red is the received signal so in a situation where you may have a fixed object at a distance and all you’re doing is seeing a change in distance you would see a waveform very much like this so we’ll talk about the the the architecture of the of the receiving system here in a minute and you’ll see why this distance is significant the shift is significant actually I think I may go ahead and skip on down to that probably we’ve been better to show this slide in advance this is a overview of what a radar system would look like very top level obviously on the right you see the digital processing portion of it and we have devices that are very focused on this application and to the left you’ll see components that we would provide for some of our radar and log devices and some of the things that dragon was talking they’re gonna be in that dark blue area although this is a very top level and doesn’t exactly reflect what dragon was talking about it’s kind of representative of what you might see in the system generally there’s some tuning control that’s creating that waveform that you’re seeing in the frequency modulated waveform and that gets modulated on top of that 77 gigahertz system so again it’s off its oscillating at a much lower frequency that drives a voltage control oscillator which feeds the resulting waveform into the transmitting section so it’s being broadcast out one of the key factors here is you’ll notice there’s an F mix signal coming off of that transmit block that F mix signal appears back now and there the received sections it’s in the receive sections where the return signal is demodulated against that that returned signal so we end up with some intermediate frequency being transmitted some base frequency being transmitted between the received channels and the eight of these that are in the rest of the system so when we’re talking about the received waveforms and the previous or this next set of waveforms you’ll see why that’s a significant factor that that changed that shift is what really gives us the information about the environment around the vehicle so what you’re seeing here now on the top is the waveform that’s being that’s external to the system being transmitted out to the the objects that are being tracked and the waveform at the bottom it’s a very simple straight line waveform in this case is the intermediate frequency that’s coming back between the receive channels and the a 2ds in the case of the stationary object the reflected signal comes back and gets demodulated into a very simple waveform so we’re now just kind of looking at a little little v-shape form that’s really telling us the information about the distance to the object because that’s that’s your flight time so very simple the equation down the bottom would just say your your your transmit speed over you’re over the speed of light by 2 and you get end up getting the distance to the object away from you so but that’s just one component that we’re interested in so distance is an aspect but acceleration of the object is another aspect that we’re interested in getting and that information would come from the Doppler shift of the object in front of you are being detected now this next chart is intended to kind of give an idea of what the intermediate frequency of that would look like but it’s an artist it’s an artificial construct because it’s impossible to have a device that’s accelerating but not changing its distance in this case this diagram is reflective of a device that is that is accelerating I think towards you but not changing its distance so sorry the detent is to communicate the the concept that you see in that intermediate frequency a straight shift off of the base so finally when you mix the two signals together this is more the real-world waveform that you’re gonna see coming back to you so to say it’s a combination of components so you see a bit of a Doppler shift pushing the received signal to a little bit higher frequency but it’s also shifted in time and not only shifted in time but the waveforms are compressing it a little bit different different rate so it’s becoming less periodic on the same at the same rate but now you see the intermediate frequency is reflecting two different resulting signatures as you ramp up you see a shift of that intermediary Qin see down off of the the baseline so and then when you see it on it’s down down cycle now you see a shift in the opposite direction so that Delta between those two shift is giving you information about the Doppler shift versus the time of flight so you’re now able to take this information put it back into this basic equation very simplified equation and your aggregate frequency shift is now reflective of all the information that’s in that in those two waveforms so now that you’ve got that signal back in we do a lot of digital basic digital signal processing that you would expect to translate that information into something that you can start to track so there’s a significant signal processing component that in the system and then there’s a significant matrix math type problem that occurs to take that information now your your signal information and put that into something that it can figure out where an object is and then take action on it so there is an interesting schizophrenia in this application I’ll call it because it is a mix of signal processing and control code traditional solutions have been put a DSP on there and then put a small processor small micro RISC processor or something in a system and have two devices well obviously it’s expensive it takes area there are lots of issues with with heat buildup in these systems and a small space and so in the future and what we’re seeing right now is utilization of architectures that combine the best of both worlds into one solution so all the manufacturers today are kind of looking at strategies to mix those two capabilities together so now when we start talking about that interface between those a 2ds and the digital components of the system there’s lots of challenges that how you bring those two together from an MCU manufacturer factors perspective putting that types of a 2ds on a single MCU is probably not very practical so one more straightforward strategy is to provide economic means of integrating external a 2ds or external Asics into the system so that you don’t have a lot of glue logic involved with bringing the two components together but you allow the flexibility to the customer and how he designs a system so typical mechanisms for bringing that analog data into an MCU would be some parallel digital interface and in fact we have defined a parallel digital interface specific to integrating high-speed a 2ds alternatively you could have bring it into an external bus and we have customers who do that but external buses have the extra overhead of a address transaction which is essentially not needed in this situation so the PDI provides specific signaling to do channel control this particular interface also provides capabilities to do frame control if you wanted to interconnect a CMOS image sensor and we opted to add those capabilities because the analog that goes into a CMOS image center is very very similar to what you would see from the digital side of a standalone A to D or an ASIC a 2d so typically in radar and people are very interested in in high precision information coming back so typically we’re seeing about 12 bit 8 a 2 D interests and so our device provides a capability of supporting any any particular size 8 10 16 up to 16 that data widths coming back into our device and up to 33 megahertz of performance typically what you’ll see from the ad DS performance I’m receiving from the received RF channel is anywhere from 5 megahertz to 40 megahertz and it depends on the customers yeah a bit I’m sorry that’s the data bit should have been a lowercase B probably up to 16 bit data wits in this case in this case you have the example is simply an 8 bit but it could be a 16 bit and it’s 16 discrete signals so we’re not doing any multiplexing this is a generic interface so it would work with a number of different manufacturers high-speed a 2 DS and CMOS image sensors it’s also very typical to have a MUX in there and actually dragon had indicated a MUX in his system and we’ve seen customers implement that in two different modes either in analog MUX in front of a single fast a 2 D or oftentimes just the digital MUX behind several discrete analog muxes in any case we can take that data however you would decide to do it and we also have channel control on our PDI that allows you to select or operate a digital MUX that might be in series with the 8 of these the other aspect of it is once you get that data into the chip you need to position it so that you can actually process it efficiently so we’ve added in our DMAs the ability to D interleave the incoming signal so typically what is done is you would read samples sequentially because you want to do some positioning information and having that data in parallel gives you an exact positioning mechanism so from from all the received channels so this particular DMA allows you the ability to build discrete buffers of interleaved data so you would see a sample from each channel and that’s what’s depicted here with us this the changing color elements so that would be some number probably you know a 16-bit data transfer going through the system our PD I actually has a capability to pack data so you will take 16-bit elements and it will build a 64-bit wide block so you can get efficient transfers through the MCU and our our crossbar architecture that’s in our MCU is very efficiently handles those 64-bit blocks because it’s 64-bit wide and as the DMA is doing the transfer and the in the buffer management of these discrete buffers the CPU is fully available for doing the signal processing component or control coded component so if you look at the basic processing tasks that are used in a adaptive cruise control implementation for example after the data acquisition the first stage would be noise reduction very commonly you would use like a fir filter a low-pass filter and you would separate the noise out often then the next step would be some windowing function that would help enhance the fft calculation that follows it’s very common to do windowing because there’s a effect of leakage that occurs when you’re sampling data and you know on the data source to become non-periodic against your sample window you’ll start to lose the energy that’s that’s it’s represented there so the the windowing function overlays another waveform and it allows you when you do the final calculations of the fft to get a more accurate reflection of the actual source frequency next would be the FFTs and we see various uses of FFTs for short-range applications you might see some very short you know 256 point maybe 512 point fft functions but as you get to the longer distances the size of those fft is increased typically and so we work with systems that are 2 and 4k points of data and then finally after you get that information translated back from the frequency domain into the time domain so you understand where the objects are in time and distance you now try to figure out so what are those responses are actually real responses and in the target detection mechanism and of those those detected artifacts which one of them actually represents an object so there’s a couple of stages in the target detection and then finally that would go into some tracking mechanism Kalman filters are kind of high-end they’re actually probably not utilized so much in in this particular application because they’re very compute intensive at least at this point and very complex to implement but we do have expertise and knowledge on Kalman filters within the organization so we do try to understand how that gets utilized but there are also other approaches to the tracking problem so I’m not going to go into an fft at this case in this discussion because there’s it would take a little bit too much time but but basically I’ll observe that the FFTs are an iterative process it has a very compact inner loop type function a summation and effect and so Mak type operations are very common in fft calculations so engines that are efficient in processing the FFTs you know would have something that’s that’s why signal processing errs our ever existed because they have very efficient Mac capabilities but these capabilities are being rolled into MC use and they’re being rolled into ours as well so we have some architectures that are efficient in handling those inner product functions and in the case of our 5500 series products there’s a what we call signal processing engine it’s essentially a Sindhi architecture coprocessor integrated fully into the EE 200 Power Architecture core so we’re able to do too Mac operations simultaneously in this in this core one of the significant things though is that we’re able to do two 32-bit floating-point Mac operations in parallel that’s something unique in a DSP world because DSP is typically operate on 16-bit functions from an FFT perspective for for a radio it’s very common to have well from FFT you’re processing one of two types of data either real type data or complex type data so you’re doing either some floating point type operation or two 16-bit integer operations in parallel so not only can we do two 32-bit floating-point operations if you decide to stay in the the 16-bit realm this Cindy architecture manages for parallel 16-bit data elements and it’s register set and these registers are 64 bit wide so effectively as opposed to maybe the more common DSPs that you would find in these applications we’re actually able now to fully process two data elements in the FFT calculations in parallel as opposed to really only being able to process one data element at a time even when they claim that there’s a dual mac on chip so I was showing an FM CW waveform in the previous slides it’s very simplistic a very triangle-shaped form but that’s not actually what’s done in this application it’s really a varying wave varying ramp on each edge of that those triangle shapes and in fact they look to get three different ramp shapes on the signal that’s transmitted out of the radar and they do that because it helps eliminate some of the artifacts that might appear that are really ghosts and not real objects so you’re essentially trying to get that same information coming back and all three legs on three legs of this this waveform so in this case you see on the far left the green red and blue ramp that is what we’re actually utilizing so it’s an up ramp and then two down ramps of different different rates so when you bring that over and you go back to the equation that we were talking about earlier and a previous slide you’ll end up maybe with some signatures that they layout over over over time based on you know the different ramps that might indicate maybe slightly different distances if you’re getting some kind of ghosted image but what you do is you now take those resulting signatures and you overlay them and a little kind of a linear regression and you look to see where all of the where there’s a response that occurs at the same point looking at the velocity versus range so when you detect something where they all overlap you you can have a fairly high confidence that there really is an object there otherwise you’ll see some kind of non-alignment of the data if they don’t intersect then you can assume it’s a a not a real object or that there’s some artifact that crept into the noise in the system that is not represented of an actual object okay so that’s actually the first half of that tracking algorithm that that rejection of the noise there’s one other aspect of it and that is so what really is an object and what two objects so the simplest mechanism that is used for determining objects is just kind of looking clustering the results and looking as at how they track over time so you’ll take sequential readings and you look at what responses actually move it the same way same rate so that’s where it’s some of the more sophisticated tracking algorithms come into play mostly matrix math the ribbon coincidentally those sim D architectures are also very effective at matrix math calculations so when we provide our MC use in this space we also provide a DSP library that includes optimized hand assembly written functions for signal processing FFT windowing fur filter functions we also have matrix math libraries to do to do basic math functions decompositions of matrices those kind of things efficiently on our sim D architecture and again they’re all written in assembly by hand so from a customer’s perspective you don’t want to be an architecture expert necessarily you want to be an application expert we have people who are architecture experts and can do at least a signal processing component we don’t do the entire application but we can enable you with the tools that would help facilitate implementing these systems in fact we have a gentleman here Patrick mentor you can also after this session discuss who’s involved with our tool support and also our modeling support working with companies like MathWorks and the customers we work with in these advance safety systems they’re very often using something like math works to help build the system code it there and the in the software were providing the libraries were providing will integrate into the MathWorks environment so that you can have a smooth transition from the model that you’re building and the and the code that actually ends up being run on our device so I have talked a little bit about architectural features that freescale offers in our MC use that address the radar application we do have a device that’s specific to supporting radar applications and it’s known as the 5561 it’s based on our ESS family of architectures that uses the e200 family power architecture core that’s designed for embedded applications the core itself the definition is for an integer power architecture core the sim D architecture that we discussed provides floating-point capabilities and it’s built right into that core so we talked about power architecture traditional classic PowerPC architecture being 32 32 bit registers 32 a number of 32 32 bit registers in this particular application there are actually 32 64 bit registers so you’re able now to operate in very wide data elements the crossbar switch and all the internal memory is constructed to support that 64 bit width so we’re transacting at each beat across the crossbar architecture 64 bit of data we also use on-chip flash for this fully integrated MCU and we have some advantages having to flash on chip in that we’re able to organize the flash array as 256-bit wide data widths which is a cache line with cache line fill with so our cache controller that’s utilized on the power architecture now is able to transact with our flash a complete cache line at each transaction it of course gets broken down and transmitted in 64 bits across the buffer but in our flash control unit there’s a buffer that mint holds that data and manages it transmission across the crossbar architecture additionally the code that’s being executed in a linear fashion we have a prefetch buffer that reads the next array of X cache line out of the flash array for positioning for the next access from the core so if you’re doing an iterative function that iterative function is going to be trapped in our cache near the near the core so it executes that one cycle access times into the cores pipeline and of course when you start to move back the linear flow after you’ve done the iterative DSP function now you’re transacting the next line which is already held in our buffer out of the slow flash into into the cache on the core side so you get very very fast operation of the flash and in fact if you look at the math it almost works out to be SDRAM speeds which is very fast for a flash and of course we have 192 K of SRAM which is architected specifically to handle several channels a double buffered data captured from an RF front end so we’re able now to do all the processing within this MCU that you would do for an ACC type application they’re their extensions are adjuncts essentially so yes they’re they’re very they’re not you know mnemonic effect really they’re this I’m really strange coding but they are they are executed within the power architectures instruction flow so the pipeline to feed that SPE engine is linear to the rest of the power architecture instructions and compilers today we’ve worked with all the compiler vendors greenhill D AB you know there’s GCC flavors of this that have intrinsic functions that allow you to access those directly I do remember GCC supporting some intrinsic sand I don’t know if that’s broadly available but I know some work was done in that ok well I’ll have to you’ve got to look to make sure that it’s in there but I do remember some discussion of that for sure Greenhills D AB code warrior they have the intrinsics built in now the intrinsics are very complex complicated to utilize they’re available to you to directly code but we also provide those DSP libraries to simplify the task because parallelization of the of these tasks is not always intuitive so for some people it can be a little bit twisted to try to look at an FF tree and figure out how I can operate on it in parallel and in fact there’s a its proprietary paper that can be made available to customers under NDA that we have that articulates exactly how that fft would function i don’t know it’s a very old one it comes from our area networking and controller systems group if you get with me I can I can arrange to have that paper so this pretty much concludes the DSP portion as I mentioned it’s kind of a reduction of what we did last year from the signal processing side because we really wanted to get you aware of the work that’s being done on the RF side because that’s very exciting to us it so it opens up a whole new area of technology that we can now exploit then bring some better value to you as customers so I feel freer you know are there other questions that we can discuss ok great well yes that’s freescale technology top to bottom yes and as I mentioned you know we’ve got a long history in wireless we have a long history in wireless and the whole ciggies see stuff is you know it was frisky it was one of the original pioneers of that types of technology and we had actually contributed some seminal papers in that area I know the engineers back then that we’re participating in some of that development so we’ve got some strong capabilities one thing that I’ve noticed in the industry that people really like to go out and talk about their ex hbt and the frequency that they’re hitting on the top top in the F max we’re we are we have a very good technology that’s equal to all of that stuff the free scale is not or motorola has not traditionally been one that really kind of bang the drum on that but now I think it’s it’s a great time for us to bring those capabilities out to technologies like 77 gig radar the 5561 is qualified silicon samples and it’s going into production at the end of this year the radar is very special so it takes a little bit extra a sophistication to to deal with that so I’d have to turn that over to Dragan I probably want to talk to you after this if you are that interested well we’ll certainly be available to answer questions if anyone likely to come up afterwards and ask questions and you’re certainly free to I know where the the first session of the day after a concert evening so hopefully we’re getting ramped up a little bit they go into the rest of the presentations today and thank you all for coming to FTF and i hope you enjoyed the experience here


Leave a Reply

Your email address will not be published. Required fields are marked *