• State TI Thinks Resolved
  • Locked Locked
  • Replies 47 replies
  • Answers 12 answers
  • Subscribers 28 subscribers
  • Views 30503 views
  • Users 0 members are here
Support feedback
Options
Options
Similar topics

This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

Multiple antennas, single RFID reader IC

Ido Kleinman
Prodigy 85 points
Other Parts Discussed in Thread: TRF7963A, TRF7960, TRF7964A, MSP430FR4133, TRF7962A

Hi,

Is it feasable to design a circuit which will contain 4 antennas and a single RFID reader chip (TRF7963A), the concept I have in mind is detecting tags in 4 different locations (~15-20cms apart) and being able to detect from which location the tag was read. Circuit should be as low cost as possible this is why 4 TRF79XX chips is not a good solution.

The concept I was thinking about is using an analog RF switch/mux IC between the 4 antennas + their impedance matching circuit and the reader IC, connecting the mux control to a microcontroller GPIOs which is also connected to the reader data interface via SPI so by firmware I can create 'continous scanning' of the 4 antennas and when a tag is detected I can determine from which mux position it was read. I am interested in the most basic tag reading (no secure/encryption, just plain ROM ID 13.56MHz tags)

Is this something that can be designed or I am overlooking some major pitfalls here?

Thanks so much

  • 0
    TI__Guru* 90906 points

    Ido - 

    totally feasible! see application note i wrote here ==> http://www.ti.com/litv/pdf/sloa167 

    and youtube video of it working here ==> http://www.youtube.com/watch?v=16VFU9n8VO4

  • 0 in reply to Josh Wyatt
    Prodigy 85 points

    I'm blown away!!! Thank you so much!

    Few questions about this design:

    1. Will it be feasable to implement also on ISO 14443 protocol? (as I wanted to use the lower cost TRF7963A IC which does not  support ISO/IEC15693 as described in the app note)

    2. Can it be modified so it can respond to reading tags from 10cm distance?

    3. Current consumption - do you think this application can work from battery power? What is the current consumption for this design? how will reading tags from 10cm affect the power consumption?

    4. Does it require any speciality RFID tags?

    Thank you so much again!

  • 0 in reply to Ido Kleinman
    TI__Guru* 90906 points

    Ido - 

    sorry for the delay - it was a busy week! 

    it is totally possible to use the TRF7963A here to read the ISO1444A tags as you say...the mux design is independent of the protocol. just make sure you tune your antennas with lower Q setting required for the ISO14443 cards in general.

    running the thing from battery power - also should be possible, but within reason of course...what kind of battery and what capacity you need is something you need to account for. 

    the distance requirement also depends on the size of the tags and the size of you antenna - what sizes are you thinking about using?

    tags are available commonly for ISO15693 and ISO14443A worldwide....i would recommend you use the largest ones you can. 

     

  • 0 in reply to Josh Wyatt
    Prodigy 85 points

    Thank you for your reply.

    Few questions/issues:

    1. I am planning to work from 2 or 4 cells of AA batteries (with a proper DC/DC). Is that something feasible?
    Do you have an estimation of the mean power consumption for the antenna muxing system with continuous RFID tag scanning? (for the TRF7960)

    2. I am planning for the maximum range I can achieve under the standard (10cm?) - is there any app-note or document I can refer regarding the antenna I should consider for the TRF796X to achieve this range? Even 3rd party solutions or products is something I can consider (as I prefer to not go into custom non-printed antenna design)

    3. Can you also linkme to an example of a tag which can fit my design requirements? (you said largest ones I can find)

    Thank you so much

  • 0 in reply to Ido Kleinman
    Prodigy 70 points
    I am a newbie to the forum and saw this thread about using multiple antennae with one reader module. I am a software engineer with minimal electronic knowledge but can work with soldering irons, IC chips and PC Board. I will need an advise on a chessboard project using the RFID technology. Also can I acquire the parts I need (antennae, etc)? Thanks for any advise.
  • 0 in reply to victor feria
    TI__Guru* 90906 points
    totally feasible! see application note i wrote here ==> www.ti.com/.../sloa167

    and youtube video of it working here ==> www.youtube.com/watch

    as far as the chess board - i implemented something exactly like that using careful choice of tag size, antenna size and Q factor of the coils - using every other column, every other row and doing intentional cross reads on the spaces which had no antenna, to reduce total number of antennas down to bare minimum - its not a beginner subject really - but if you want help - let me know - and find some access to a network analyzer with cal kit if you can - it will be critical to your success and i can explain to you how to use it well enough to accomplish what you will need to do.
  • 0 in reply to Josh Wyatt
    Prodigy 55 points

    Hi Josh,

    I am trying to implement RFID for a chess board for my senior project in my computer engineering degree. Using the antenna mux seems like my best option. A big part of my problem is that I am new to RFID. My goal is to have a system that detects when a piece is lifted from the space and then when it is returned to another space so I can track board layout throughout the game.

    Any advice you have on where to start would be greatly appreciated!

    Thanks for the help!

    Jerome Selig

  • 0 in reply to Jerome Selig
    Prodigy 70 points

    If I have the parts and circuit diagram  and some instructions on the workings of the system then the only challenge would be programming.  First I need to acquire the parts.  How much are we talking as far as cost is concerned.

  • 0 in reply to victor feria
    TI__Guru* 90906 points

    Jerome / Victor -

    i suppose you are working together?

    you want to implement something like this? (maybe without the robot? two human players?)

    this was done with antennas on each square - later i figured out how to do row and column scanning, similar to the way keypads work - then i removed every other column and every other row and set up system to do intentional cross reads on the spaces where antennas were not physically - this means the hardware becomes pretty simple, but some details have to be paid attention to - how much time do you have on this?

  • 0 in reply to Josh Wyatt
    Prodigy 55 points

    Actually Victor and I are not working together, but is cool to see that I'm not the only one trying to deal with this issue. 

    This picture does seem to be what I'm trying to do but I will have 2 human players. Since this is my senior project I will have until approx. May 2016 to finish the whole project, but this RFID problem will not be the only feature that I have to implement so I need to address this as quickly as possible. I'm really interested it trying to do this with the intentional cross talk but since I am unfamiliar with RFID implementation I don't want to bite of more than I can realistically handle within my timeline.

    Thanks for the advice,

    Jerome

  • 0 in reply to Jerome Selig
    TI__Guru* 90906 points
    how big are the squares on your chess board?
    for reference, i found this on google ==> www.chessusa.com/.../chess-board-size.html
  • 0 in reply to Josh Wyatt
    Prodigy 55 points
    I am still in the proposal stage of this project so that size is flexible. I think my goal is to make the spaces be 2 inches but again that is flexible if other sizes will simplify RFID usage.
  • 0 in reply to Jerome Selig
    TI__Guru* 90906 points

    see attached - i just made this for you - don't want to take away your opportunity to design/learn/experiment your own thing, but this should give you a good base to start with. in this case i would choose 1.5" spaces, as this gives you total of 12" x 1.5" to cover, which is doable if you use our 24.2mm tags for the pieces

    reader IC can be TRF7964A or TRF7962A and use MSP430FR4133 might be good choice (or similar) and Peregrine Semi SP4T adn SPDT switches work well in this type of application.

    since you have plenty of time, the thing to experiment with here is the width of the antennas themselves as it relates to coverage of the adjacent row. What you are trying to do is intentionally cross read and keep track of what you read from which antenna so if you read a tag from four antennas, you know its located on the square in the middle of those four or conversely, if you only read a tag on two antennas, that tag will be where they intersect.

    make sense?  

    RFID Chess Board.vsd

    RFID Chess Board.pdf

  • 0 in reply to Josh Wyatt
    Prodigy 55 points

    Wow this makes all of this so much clearer! Thank you!

    One question is that for my application my board has actions to complete between the time the piece is picked up and when it is placed in a new spot. Assuming I wanted to reduce the vertical range of the antenna (if the piece is lifted straight up) to less than the size of each space would I have to simply use a lower power antenna for each column and row opposed to the setup that you showed.

    It just seems like the player would need to lift a piece farther than 1.5 inches off the board for the sensor to realized that it has been lifted and if I could reduce that distance that may be a little better.

    I hope this was somewhat clear.

  • 0 in reply to Jerome Selig
    TI__Mastermind 26410 points
    Jerome,

    Range is effected by many factors. The tuning and power or the reader antenna are important, as well as the tag antenna geometry, orientation, and technology type. For this application, I would recommend using ISO15693 tags, but you can still experiment with difference sizes, etc.

    From the reader side, to reduce range you can use 1/2 power mode or tune the Q of the antenna's a little lower.

    Typically, having to much range is not much of an issue. :)

    Thanks,
    JD
  • 0 in reply to JD Crutchfield
    TI__Mastermind 26410 points

    Hope you guys are trying to do something like this?

    Once you get it working, throw up a demo on you tube and post it here.    

    Thanks,

    JD

     

     

  • 0 in reply to Jerome Selig
    Prodigy 70 points

    I tried doing this using an Arduino and an 8x8 led matrix IC that controls 64 LEDs.  Instead of LED  I used a relay  IC that connects each antennae using only one reader module.  The software controls which antenna is connected.  The reader then transmit the data read back to the Arduino and relayed to a PC which analyzes the process.  I  will sow my prototype later post.

  • 0 in reply to Josh Wyatt
    Prodigy 70 points
    Is this something you can share?
  • 0 in reply to Josh Wyatt
    Prodigy 70 points
    In my case I use 2-1/4" squares which is standard tournament size.
  • 0 in reply to Josh Wyatt
    Prodigy 70 points
    In my opinion crossreading makes it unreliable since in actual play humans do not usually place the piece on the squares center
  • 0 in reply to victor feria
    TI__Guru* 90906 points
    the intentional cross reading makes it less complex - its same reliability as reading with each square or each row and column.
    if you have a doubt, you can always also use RSSI values (requested from TRF796x or TRF797x device as final vote, too
  • 0 in reply to Josh Wyatt
    Prodigy 70 points
    can we have the specs on the antenna on that board?
  • 0 in reply to victor feria
    Prodigy 55 points
    It sounds like that is very dependent on the board that you are using and that we will have to experiment and tune our antennas to get the range that works with our individual board implementation.
  • 0 in reply to Josh Wyatt
    Prodigy 10 points
    Hi, I am new to this community but have been using this thread for a while as reference to a potential project I am working on and love where it is going so far. I was wondering if you had any comment on if this would be possible to run through an Arduino or Raspberry Pi type device? I was hoping to build something like this-
    www.youtube.com/watch
    I've seen this unit before and they say they use an open source tangible table system with an IR camera underneath, but I think a RFID system could be more compact and cost effective. I think using your design of the RFID Chess Board is ideal as only an X and Y axis would be necessary and not necessarily need to be linked to each other (Y would alter the sound one way, X a different way). I'm pretty Green to programming and all of this, but have a very strong community of very capable people who would love to help with something like this if we had some direction. Anything helps! Thank you for what you have done thus far!
  • 0 in reply to Cecil Wilcomb
    TI__Guru* 90906 points

    Cecil -

    yes - if you see this app note - this describes such a system which i implemented completely at least twice and then have a third version of it cut down to just the four.

    you can make a "game board" like that - just have to keep it reasonable on the # antennas - if it gets to be too many, then you can revert to a row and column approach on the antennas or an every other row, every other column approach - both of these last two require patience when designing and prototyping as you have to pick a transponder and marry it to the antenna tuning / Q, relative proximity to the other antennas, etc. its a little tricky for a new person, so i would stick with the basic approach to prove your concept out.

  • 0 in reply to Josh Wyatt
    Expert 1745 points
    I am wondering about the RF switch you use in your application note. You are using the Peregrine PE42440 for switching the 13.56MHz antennas to the reader, but the PE42440 is specified as a 50-3000MHz switch. Are there any issues as the antenna frequency is lower than the switch's lower limit?
  • 0 in reply to J.Schweiger
    TI__Guru* 90906 points
    if you see the end of that app note where i characterized the loss on the switch at frequency which correlated with what they have in their datasheet. I also contacted PE and they reviewed my use case and saw no issues at that time.
    if you look at the way the curves go on page 5 of the datasheet www.psemi.com/.../pe42440ds.pdf
    you can see its OK - now i do see it looks like they updated the datasheet with a figure 12 i do not recall being there before.
    it indicates you need to reduce VDD to the PE device to 2.65V and then if operating at full power out of the TRF79xxA (+23dBm) you would be over what that curve shows.

    i never saw any issues with blowing up parts when i used them - but i also did not do a great # of these setups (maybe 6 or 8) - anyway - technically i would still have same opinion the switch is fine in the application - but you would want to reduce that VDD into the switch and maybe either run the part @+20dBm (register setting of 0x00 would be 0x31, if running at +5VDC in, +3.3V operations should be fine already) to be on the safe side in accordance with this information.

    hope that helps.
  • 0 in reply to Denia
    Prodigy 20 points
    Can you explain why the top row, for example, would not also be picked up by antenna 5? Antennas 6 and 7 span three rows, but 5 only spans two?

    Also, how are you supposed to identify the top-right corner (which is blank) ?
  • 0 in reply to Stephen
    Intellectual 260 points

    Posted for discussion and the right to life my ideas.Who that can prompt

  • 0 in reply to Stephen
    TI__Guru* 90906 points
    you would get a read from antenna 5 and 4, if you set the size and Q correctly to match the size and Q of the tags being used, to see the top right corner
  • 0 in reply to Josh Wyatt
    Intellectual 260 points
    I'm from Russia and it is difficult to understand the meaning in the translation.Identification of shapes based on RFID tags ,and chess cells at the intersections of the antennas.How to determine the cell name, if all options of the interaction of the antennas 4 and 5 is already in use?
  • 0 in reply to Josh Wyatt
    Prodigy 20 points
    Okay, but the square below and left of it is also 4 and 5, so then how would you differentiate those?
  • 0 in reply to Stephen
    TI__Guru* 90906 points

    so the algorithm is switching through the antennas - the antennas are to be setup to do intentional cross reads on the squares (actually rows or columns) that do not have antennas - that is the advantage to doing it this way (less antennas) - as the alternative is to have antenna on each square or antenna on each row and column - it does require some knowledge, ability and patience to tune the antennas correctly and choose the correct tag size to match the system, but it works.(like this, i think you understand that - just showing here a simplified view of some of the flux lines that would be generated. 

    to answer you directly - i see what you mean - in that case you would have to rely on UID and RSSI value returned to determine where it was. When i worked on this i considered odd number of antennas to make it work - you could add a dummy column and row to get same effect

    image on the right is with antenna on every column and row, image on right is reduced # of antennas as we are discussing, but with 15 x 15 array, as that is what i needed to accommodate with what i did. 

  • 0 in reply to Josh Wyatt
    Intellectual 260 points
    maybe 2 coils on an empty chess cell to indicate any 2 antenna.Don't ignore my posts, I'm just learning
  • 0 in reply to Josh Wyatt
    Prodigy 20 points
    Hello Josh,

    your plan is very helpful and i begin to understand, what the components are doing.

    But i am software expert (cryptology specialist) and my experience with those type of electrical parts is 20 years old (i had contact at study times last). sorry, but due to this i am unable to follow you in detail.

    i would like to build such a board with my kids, so i would order those parts, right ?

    TRF7964A = TRF7964ARHBR = store.ti.com/TRF7964ARHBR.aspx
    MSP430FR4133 = MSP430FR4133IG48 = www.ti.com/.../samplebuy
    SP4T = PE423641 = www.psemi.com/.../pe423641
    SPDT = PE42359 = www.psemi.com/.../pe42359

    but what do you mean with "tuning cuircuit" exactly ?
    what do you use as antenna ? simple cable or something special (copper / special product)

    your help would be very important for me.

    maybe you can understand my problem. i will have my focus on the software development and show the kids the different ways to communicate with those technical parts (f.eg with changing the software for playing checkers).

    Kind regards,
    Kim
  • 0 in reply to Kim Johansen
    TI__Guru* 90906 points

    Kim -
    IMHO - this is not a beginner project from any point of view and all these devices are surface mount, so unless you have experience with SMT assembly (and the correct soldering iron(s) and a microscope with which to mount them correctly, plus access and experience with a network analyzer) I would not attempt, if I were you - you would need to be an expert in these fields to build this in the first place, let alone teach someone else how to do it.

  • 0 in reply to Josh Wyatt
    Prodigy 20 points
    Hello Josh,

    thank you for your reply.

    of course i know its not basic; but i am able to build technical projects including soldering iron(s) etc. i build complex projects with raspberry pi , arduino and bananapi (usage of gpio interfaces inkl. SMD build etc). for most of them i used my instrument with special clamps and magnifiers. they all are working and brought me new skills. but my focus is software developing and mathematical knowlege.

    i would like to learn with this project, test with the antennas to find the best solution etc. ... but simply i am not an expert with chips and selection of special (matching) components.
    that's why i am asking: i would like to try it ... but buying wrong components or trying with wrong antennas would never work. i will learn and try; but i will not throw money out of the window for nothing (f.eg. wrong chips). with the correct materials i will find a working solution ... if your interested in, i can send you the assembled product or pictures, so you can see what somebody can build our of your informations ;)

    private people coming to me and asking for "how to simply crypt or decrypt". nearly all of them are no mathematicians or it specialist like i am ... but i like their interest in my skills. so i am not asking them to study mathematics or software engineering; i am giving them f.eg. my piece of source code and they can try to fit it to their needs. mostly they need help again, but that's no problem. after all and some help to find the correct way, they were able to build a piece of software (nearly) by their of and they have learned a lot. sometimes we got nice and long termed contacts out of such requests. in this case i am the newbie asking for some help, pls.
  • 0 in reply to Josh Wyatt
    Prodigy 30 points
    Hi Josh,

    I have an application similar to the chess board except it has 16 squares. I was thinking of using 16 antennas (one under each square) and mux it to TRF7960. My vertical read range requirement is 5cm and on xy plane is a square of 7cm by 7cm. RFID tag is 1cm diameter.

    Not sure whether the tag size and reader antenna size would work. Can you advise?
  • 0
    Prodigy 50 points

    Josh,

    I have been fallowing these post and they have been very helpful. I have a question could this type of chess board be set up in an outdoor environment on a bigger scale. 

    I would like to know if it would be possible if I could measure the distance of multiple objects in an area anywhere from 10 x 10 ft to 20 x 20 ft using RFID tech. For example, Say I have a golf green (mini golf) that is 10 x 10 ft and I would like to know the distance the ball is from the cup (pin) when the ball is on the green or in that area. I would need that information relayed back to me in real time. But also if there were multiple balls on the same hole, I would need to know all their distances as well and if they were to move (if they were hit be another ball).

    after reading through these post i figured  I could place a HF chip in the ball and possibly create a grid of antennas and readers. Like a chessboard with an antenna under each square and a tag on each chess piece. The reader would fire up each antenna in sequence. When it reads a tag, I know the ID of the tag and the object associated with it. Since the reader identifies the antenna that read the tag, I would know each piece’s location. I believe the same could work with balls on a green? I would like to know is if this is possible to create outside? What types of antennas and readers would I need to use? 

    Note: The antennas would possibly under fake turf. i also encourage other suggestions on how this could be made possible another way.

  • 0 in reply to Kelvin Sealy
    TI__Guru* 90906 points
    this would be super expensive if you even tried to experiment with HF as you would need high power readers and antennas from Feig or similar vendor - for this kind of distance. Even for UHF RFID it will not be cheap i think. What is your budget?
  • 0 in reply to Josh Wyatt
    Prodigy 220 points

    Josh, thanks for this interesting thread. Can you explain in more detail on how the above was designed?

    Not seasoned on RF circuits so have many questions.

    1) is the reader circuit operating @ 13.56 mhz? Read that 125khz may be more practical to intentionally force the SR (short range) pickup of the tags?

    2) in the above case, when operating, does each reader antenna pickup only a single tag or does it pickup also nearby tags? if multiple tags are picked up, how did you cancel the surrounding tags from your logic? would be great to have a reader IC at these frequencies with a RSSI register to then isolate and select the tag with the strongest ping value.

    3) how close can the antennas be placed without having a negative impact on the operation? thought to place a single antenna inside each 2x2 square. welcome more detail on you were able to skip columns / rows as still not clear on the concept. can you share the PCB antenna layout for the above project? This will assist to reduce the BOM.

    Due to the latency of the RFID pickup, will consider to use 1:4 or 1:8 mux as a max in this proof of concept. We would like to start with perhaps a 4x4 grid or 1x8 of antennas and use the mux to cycle the antennas at varying times so that they do not impact each other when the respective square antenna is enabled.

    Would a RF balun or similar be more suitable for an intentional short range pickup of the RFID tag? That is, to not pickup other tags that may be sitting along the perimeter of the PCB antenna and want the tag that is center to the antenna layout.

    Thanks!!

  • 0 in reply to Mon2
    TI__Guru* 90906 points

    1) 134.2kHz HDX, 13.56MHz and UHF RFID technologies all have their place in the world. 

    see attached for a short primer on the three (from an application specific note written about 8 years ago) 

    RFID Selection Guide for Waste Management Applications.pdf

    2) each technology type has their own protocols to handle arbitration of multiple transponders in the field. it would be time consuming to elaborate fully here - but you can read the ISO standards that govern the techs to learn more. To summarize LF - usually a "gate mode" is used, although anti-collision does exist, because of the speed of the air interface, it takes much longer than the other two main techs to perform that. HF tech has multiple protocols which use that band. ISO15693 is the longest range of these and it (as well as ISO14443B) uses a slotted approach for anti-collision. Typically, when looking for tags, the reader would be sending out single slot inventory command (for ISO15693) or REQB/WupB to activate - if and when a collision was detected, the reader firmware would switch to sending out the slotted version of those, then arbitrate. ISO14443A uses anti-collision as part of the activation sequence to begin with. FeliCa uses a slotted approach, but it is a time based one in which the card selects a time slot randomly, based on value sent from the reader during the Polling command. UHF RFID uses a random protocol for same. 

    3) the idea of doing rows and columns came from previous work i had done on keypads. it was superior to using antenna for each space because it used less antennas (cheaper / easier to route) - then since i knew a few things about magnetics - thought to simplify further by removing every other row and column. it took some effort and experimentation with different tunings, tag sizes, Q values, etc. to get it right. 

    4) here is an image of a board i made to start with - you could do same

     

    Lastly - a balun is generally not needed in HF RFID...this would be more something that is used for HAM radio applications

  • 0
    Prodigy 10 points
    Hi Josh,

    I'm an engineering student currently working on improving a system similar to that of the chessboard. I am trying to improve an already built design, where there is one antenna per reader with 18 of these modules in total (3x6). However, the cost really scales up when a larger area needs to be covered; thus, I would like to try the multiple antenna approach that you have implemented. As this project uses 125kHz tags (which can not be changed), I am wondering if there are any RFID readers you may suggest that can implement the design you posted in www.ti.com/.../sloa167.pdf.

    In addition, after reading through the document, I am a bit confused to whether you used two RF switches for each set of antennas, one for antenna input and the other at antenna output or were the switches only used at the antenna input since the other end of the antenna connected to ground? Sorry, I'm a bit lost here about your antenna schematic as I am only using custom built coil antennas at school.

    Thanks,
    Becky
  • 0 in reply to Becky
    TI__Guru* 90906 points

    Becky - 

    I would never try this with 125kHz or 134.2kHz - the read range at close distance would require anti-collision which is too slow (because of the data rate possible with low frequency) IMHO for the application.  

    The high frequency reader i used here was single ended, not differential, so the antennas were all matched to 50Ohms and i switched

    if you see figure 4 on page 5   

    you will see i cascaded the switches and also you can see the truth table on page 6