RFID Read Ranges Revisited

John R.

I had posted earlier about using RFID on a model railroad.· At that time (pre prototype hardware testing) the concern seemd to be limiting the read zone because the tags could get too close together, etc.· (there are still concerns about overlapping readers, but I haven't gotten that far yet).

In reality, now that I have BS2, Readers and tags in hand, the problem is the opposite.· I can read the tags, and read a "string of tags" very close (even overlapping for the large "world tags").· What I can't do (yet, by admitedly crude attempts) is read a tag once I put it inside a plastic box car (.5 dia x about .1 thick "button").· If I hold the antenna right up to the side of the box car, I can get an intermittent read.· This won't hack it.

I'm sure larger tags will help.

I just thougth I'd throw this out, hoping someone may have some suggestions on getting this to work better.· I can think of all kinds of potential sources for problems, and havn't started tracking them down yet (and am at work now).

Does anyone have any insight that might be of help?· Any other sources of info (for a hobbyist with not much more than basic understanding of the basic concepts involved)?

Thanks;


▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
John R.

8 + 8 = 10

Newzed

The tags in the box car - are they standing up or laying down?

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Sid Weaver
Do you have a Stamp Tester yet?
http://hometown.aol.com/newzed/index.html

·

Bean

How far away can you read the tag if it's NOT in the box car ?
A plastic box car shouldn't make much if any difference.
Is there any metal in the box car ?
Bean.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
"SX-Video·Module" Now available from Parallax for only $28.95

http://www.parallax.com/detail.asp?product_id=30012
Product web site: www.sxvm.com
Available now... SX-Video OSD module $59.95 www.sxvm.com

Life is NOT a box of chocolates.
Life is a trip, and 99% of the time you get where your headed.

·

John R.

These were lying down flat, and the antenna was vertical, RIGHT next to the car in order to get the intermittent reading.

I tried the antenna flat, under the car, with absolutely no reads. I suspect the "weight plate" in the car may be the issue there. I have by means exhausted alternate orientations and configurations.

I've just started working this problem, so any ideas or pointers are welcome.

I'm also looking at some of the "capsule" type tags, hoping that I can mount these under the cars, and closer to the reader antenna (ideally mounter under the track).

I will be working more on the "inside the box" (not box car, but figurative box) solutions (distance from antenna to tag, tag/antenna orientation, proximity to metalic objects, etc.) tonight and over the weekend.

I'm also interested in any experiences "outside the box" (different antenna configurations, power boosting, etc.) that might be helpful (or knowing what NOT to do).

I've tinkered with electronics in the past, but nothing this "serious". I'm more of a "software guy". I have the aptitude, but not the experience, and am still on the steep part of the learning curve.

This application is currently for my personal use, but I do have delusions of grandeur of a commercially viable product (feed the data to a PC interface).

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
John R.

8 + 8 = 10

Chris Savage

John,

·· Try standing the tags upright...They won't work as well when perpendicular to the reader.· They should be paralel to it for best results.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Chris Savage
Parallax Tech Support
csavage@parallax.com

John R.

Actually, the antenna will normally be "flat" and under the car (so a laying down tag would be "optimal"). The "Crossed" configuration was just something I tired to get the antenna closer to the tag last night. If I hold the tags in my hand, things work much more reliably. I will be delving into the actual construction of the box cars, and alternate tag locations tonight and over the weekend. This also gets easier for the larger scales (bigger tags, more distance between them).

I suspect the "magic bullet" is somewhere to found on tag shape (capsule, button, etc.) and location/orientation. I probably will need to mount the tags externally, on the bottom of the cars. Doing this in an aestethically acceptable manner may be a challenge.

Luckily, I have a bunch of processes running at work, and have time to ponder today...

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
John R.

8 + 8 = 10

Newzed

John, I have been this route.· Tape the tag to the inside of the box car in a vertical position·so it is "broadside" to the vertical antenna.· It should read perfectly in that position.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Sid Weaver
Do you have a Stamp Tester yet?
http://hometown.aol.com/newzed/index.html

·

John R.

Sid;

This is fine for testing, but having 4 x 4 antennas all over the layout doesn't seem too practical. (not to mention things like flat cars)

Did you ever get a reader working from "under the track"?

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
John R.

8 + 8 = 10

Newzed

Never did.· The weight plate at the bottom of each car is too much interference.· I wound up mounting the reader antenna vertical about 2 inches from the side of the train path.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Sid Weaver
Do you have a Stamp Tester yet?
http://hometown.aol.com/newzed/index.html

·

John R.

I'll let you know if I have any success. My current thought is to get one of the "capsule" type readers under the weight plate, like in a brake cylinder or attached to the truck. I may start "ugly" (hang it in a loop of tape) just to see if I can get it to function, and then go from there.

How many readers did/do you have on the layout?

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
John R.

8 + 8 = 10

SPENCE

JUST SOME IDEAS. LAYING THE TAG DOWN FLAT OVER AND CLOSE TO THE METAL FRAME IS GOING TO ABSORB RF ENERGY AND MAKE THE CARD READER INEFFECTIVE. I CERTAINLY AM NOT A RF ENGINEER BUT I THINK YOU MIGHT TRY SETTING GOING THRU A RF TUNNEL. (SOME METAL SHIELDING WHICH WOULD HELP CONCENTRATE THE RF IN A SMALLER AREA). 136 KHZ SIGNALS TEND TO SPREAD OUT EASILY DUE TO THEIR WAVELINGTH. YOU MICHT TRY A METAL FOIL TO CREAT A SMALL TUNNEL. TRY SEVERAL DIFERENT TYPE OF MEDIA., FOIL. TIN LIKE GALLON SIZE TIN CANS THAT FOOD COME IN. YOU MIGHT GET SOME FROM A RESTRAUNT. MAYBE EVEN SEVERAL IN A ROW FORMING A TUNNEL THROUGH A MOCKUP MOUNTAIN. SOLDER THE CANS END TO END

THE READER PC BOARD ANTENNA HAS TO BE PARALLEL WITH THE RFID TAG.......

73
SPENCE
K4KEP.

Post Edited (SPENCE) : 10/27/2005 5:09:46 PM GMT

Newzed

John, I only had one reader - I was just playing aroung with it.· I finally decided it was more important to know when the lead engine passed a certain point than it was to know about a certain car, si I installed four IRE systems to detect the engine at points A, B, C and D.

One other thought - the track is nickel-plated steel, which is quite magnetic.· If you put your reader completely under the track, I think it will still interfere with the performance of the reader.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Sid Weaver
Do you have a Stamp Tester yet?
http://hometown.aol.com/newzed/index.html

·

John R.

I'm working in N, HO and/or On30, so the rail is a brass alloy with high nickel content ("nickel silver"). I'm more concerned about possible interference from the power and DCC signals going throught the track. Right now, I'm taking that all out of the equation for now, and starting with just rolling the car over the antenna. (crawl, then walk, hopefully run latter).

In terms of practical application, this is overkill, (unless you're modeling modern equipment and technology) but my railroad hobby is driven by a disire to see what technology can do, and head toward automated operation.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
John R.

8 + 8 = 10

Danny

How about disguising the reader as a billboard next to the tracks?

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
"Never create anything you can't control"
"The amount of intelligence on the planet is fixed... the population is growing"

Bean

Maybe you tried this, but what about putting a bar code on the bottom of the car and "read" it as the car passes over an optical sensor ?
Just trying to throw out some other ideas...
Bean.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
"SX-Video·Module" Now available from Parallax for only $28.95

http://www.parallax.com/detail.asp?product_id=30012
Product web site: www.sxvm.com
Available now... SX-Video OSD module $59.95 www.sxvm.com

Life is NOT a box of chocolates.
Life is a trip, and 99% of the time you get where your headed.

·

John R.

Bean;

Catching up on an earlier post, I can read the button at about 2" outside the car, and the capsule just at a little over 1". There is a metal plate in the car for weight. I suspect this, and/or power and DCC signal in the track may be the problem.

I just tried taping a capsule to the bottom of the car, and this looks to work so far, but aesthitically(sp?) won't cut the mustard for some modelers (in the small scales anyway). This was not over powered track. That will have to wait until over the weekend. I've also got some smaller capsules coming, and I hope to fit these between the wheels in the truck. There is so much "non-scale" stuff happening there that the appearance will not be as much of a problem. It also gets the tag closer to the reader.

I thought about bar codes, but there are aesthetic problems there too. I haven't ruled it out, but am plunging down the RFID road first.

I also haven't found a bar code reader as cheap as the RFID reader hear. I realize that I could probably work something out in the STAMP for a crude bar code reader, parsing light and dark, but the complications of speed variations (or stopping/reversing) complicate that.

Keep the ideas rolling...

As a side note, I am having a blast working with this stuff, and the help, ideas and interest shown here have been fantastic.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
John R.

8 + 8 = 10

John R.

Danny said...
How about disguising the reader as a billboard next to the tracks?

Danny;

In some situations the disguising may work, in others (e.g. circa 1900 mountain railroading) it gets harder. Also, in "N" scale (1/160) the 4" reader is over 50' tall. Quite a billboard...

Depending on how one wanted to use this, it might still be an option. Under the "delusions of grandeur" heading, I'm trying to come up with a "universal" solution.

If it comes down to this, I think it (disguising) could still be done. Maybe not as a bill board, but in a building, the side of a cliff, inside a tunnel, etc.

The gears are grinding away...

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
John R.

8 + 8 = 10

Chris Savage

Hmmmm, I seem to recall the barcode reader/RFID thing being hashed out, but I don't know who the OP was.· In any event, trying to offer an alternative, if the bottom of the cars are metal that will adversely affect the operation of the RFID Reader.

If you turn the ID Cards (Depending on their size) at 45 degree angles, they will be equally effective to a degree whether the reader is on the bottom or side, barring any interference from metallic objects (The tracks themselves could affect this).

As for having some RFID Readers sticking up around your track, couldn't you disguise them with plastic buildings or even paper/cardboard scenery/buildings?· I hope you know what I mean/refer to?

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Chris Savage
Parallax Tech Support
csavage@parallax.com

John R.

I know there was some barcode/RFID here in the sandbox on my original thread a few weeks ago, before I had hardware in hand.

Disguising will work fine for the bigger scales, but is a definite problem for the smaller (as mentioned, the reader is over 50 scale feet for N gauge or 1:160). Trying to hide a row of 5 or 6 of these at a yard entrance could be difficult.

I have enought ideas to work on for a while. My own railroad is with larger scale (1:48) so life is easier on many fronts, including the ability to use a larger tag. It is also a narrow gauge short line mountain railroad, so the opportunity to hide the readers is definitely a possibility. Most of the current angst, trials and tribulations are with trying to get the technology working for N scale. If I can get it working there, it can work "anywhere", at least in terms of Model Railroading (I'm not worried about Z scale.)

Hopefully, I will have a chance to do some playing -er- more serious testing - over the weekend, and I'll let you know what I've found out. Plans include under body mounting of the tag, removal (at least for testing) of the metal wieght plate, and various mounting methods for the antenna.

I have not totally discounted the thought of trying to create my own antenna, and/or some type of "directional" shielding or "funnel" as mentioned in this thread above.

While I may be seeming to discount some of the ideas put forth, please understand that is not the case. I have some pre-conceived ideas about how I want this to work out, and will be proceeding down that path as long as possible. Once all thoughts have been worked on, I may end up coming back to "reality". Right now, I'm still trying to find define the envelope and see what I can make happen. Thanks to everyone for all the help, and by all means, keep any ideas coming.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
John R.

8 + 8 = 10

kelvin james

Maybe not what you want to get into, but what about " active tags" ? They are a powered tag, with a 3v coin cell battery, and have a long range. It could be read anywhere on the track.

kelvin

John R.

If they are affordable, and the readers are affordable, I'm interested. Most of the "good stuff" seems a bit too expensive for the intended application.

Potentially a model railroad could have several (to dozens) of readers, and a tag for every car (over 100 would not be unusual). (Model railroaders seem to collect rolling stock and locomotives the way Amelda Marcos (sp?) collected shoes.)

At $39.00 and down for the reader, and in the $2.00 range (or less in quantity), this 125 kHz passive read only stuff seems about as inexpensive at it gets.

Size is also a factor. Putting something with a button battery in a box car is one thing. Where to hide it on an N gauge flat car is another matter.

If you have a source for another affordable technology, please let me know.

As a minor progress note: by using a tape loop, I was able to hang a 13mm x 3mm dia "capsule" tag just above the rail head. I had the reader between the bottom of the ties and the cork roadbed. I was able to read tags, as long as I kept the speed down. Progress.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
John R.

8 + 8 = 10

SPENCE

MAYBE A RETHINK OF STRATEGY IS IN ORDER. SIZE ON THE N GAUGE IS GOING TO DO YOU IN.

NEW STRATEGY. RAM MEMORY BANK CABABLE OF HOLDING ID OF EACH CAR, OF EACH TRAIN, OF EACH TRAIN DIRECTION, OF EACH TRAIN LENGTH, OF BLOCKS ENTERED OR OCCUPIED.

SENSORS DETECTING EACH CAR AS IT PASSES. MORE HORSEPOWER IN THE COMPUTERS AND LESS IN THE IR OR OTHER DETECTION SYSTEM.

MASTER DATABASE OF THE ENTIRE LAYOUT.

YOU MIGHT EVEN THINK OF USING A PC FOR MASTER MEMORY AND CONTROL WITH PARALLEL PORT IO. THERE YOU HAVE CAPABILITY OF 3 PIO'S AND THEY CAN BE MULTIPLEXED TO INCREASE THE SINGLE BITS AVAILABLE. ON DOS, UP TO 4 SERIAL I/O AND IF YOU USE LINUX DOZENS OF SERIAL PORTS. EACH SERIAL PORT CAN BE EXPANDED INTO MULTIPLE PARALLEL PORTS ALSO.

MOST OF THIS COULD BE WRITTEN IN GW BASIC. BELIEVE YOU ME. GW BASIC ON A HIGH SPEED MACHINE IS SO FAST. I GET 1000'S OF LINES OF DATA BASES PROCESSED IN HALF A SECOND AND DONE.

THIS WAY YOU MIGHT EVEN BE ABLE TO USE THE INEXPENSIVE SX/B AS SMART DISTRIBUTIVE PROCESSING.
IN HIS BOOK "BEGINING ASSEMBLY LANGUAGE FOR THE SX MICROCONTROLLER", AL WILLIAMS REFERS TO IT AS VIRTUAL PERIPHRERALS. PROBABLY IN MANY CASES THE SX CHIP WOULD NOT EVEN REQUIRE A CRYSTAL OSCILLATOR FOR SLOW SPEED OPERATIONS. IN GUNTHER'S BOOK HE SAYS THE SX CAN RUN AT 33, 128KHZ AND 1 AND 4 MHZ WITHOUT A CRYSTAL OR RESONATOR.

NEXT WITH DOZENS OF MICRO AND I MEAN TINY PROCESSORS ALL OVER THE PLACE (IN MANY CASES 1 SINGLE CHIP WITH ALMOST NOTHING ELSE) . NOW YOU NETWORK YOUR COMMUNICATIONS BETWEEN THE VP'S AND THE MASTER. OR EVEN USE SOME OF THE VP'S AS MINI MASTERS.

READ FOLLOWING WITH SNAPSCII FIRST.
http://www.hth.com/snap/
http://www.hth.com/filelibrary/pdffiles/snap.pdf
http://www.hth.com/filelibrary/pdffiles/snapscii.pdf

BEAN, YOU WOULD LOVE THIS I THINK.

73
SPENCE
K4KEP

John R.

Spence;

Basically, the intent is for the RFID scans to be sent to software running on a PC (sorry, windows interface . Storing the trans, cars, order of cars, etc. in RAM would work for "virtual tracking", but that isn't what this is about.

The whole concept here is not to know what "should" have happened, but what actually DID (or IS) happening.

Examples of where trying to track things "virtually" break down are cars becoming uncoupled, and/or the "wrong" car being dropped at a siding.

The fact that RFID requires no action on behalf of the operator is another advantage. If you tried to hold the consists of all the trains, sidings, etc. in RAM somewhere, you've now put extra load on the operators in having to say they dropped car xyz off at siding 23, etc.

The other concepts you mention (multiple units networked, and other "periphial" devices) are planned as future developments, for additional automation and signalling tasks, from basic control of RR lighting, control of annimated scenes, etc. I appreciate the thoughts for future development.

Right now, my focus is on this application, which I will either get working, or beat to death. My simple test last night with the tape loops showed me that it can work, now I just need to optimize the tag/antenna configuration/orientation, etc. Maybe in the end, it won't work, at least on the smaller scales, but I'm having a lot of fun in the process!

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
John R.

8 + 8 = 10

Beau Schwabe

John R.,
I think I had mentioned the use of bar codes in the previous discussion....

http://forums.parallax.com/showthread.php?p=548606

You had mentioned that you did not want a barcode·for "aesthetic problems".
Consider that the "barcode" would be a label on the underside of each car.· This
could be a high contrast B/W label or reflective·label.· With an IR transmitter/receiver
located on the track and adopting a coding scheme below, you could determine
speed,direction, and which car passed over the sensor with a very small aesthetic
blemish for the IR sensor on the track.
·

SSSS  D0  B0  B1  B2  B3  B4  B5  B6  B7  D1  SSSS

 SSSS = Sync Bit
D0-D1 = Direction Bit
B0-B7 = 8-Byte Data
 
SSSS is always "1" and can be measured to determine "speed" and/or baud rate.
Typically this is twice as wide as a regular bit width.
 
D0 is always "0"; D1 is always "1" ... If the first bit following the "Sync Bit"
is 0 then the next 8 Bits are LSB.  Likewise if the first bit following the
"Sync Bit" is 1 then then next 8 Bits are MSB.  This way you can determine
direction and properly reconstruct the data.
 
B0-B7 is the actual data that identifies each car.  (Could be expanded if 256 is not enough)

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Beau Schwabe

IC Layout Engineer
Parallax, Inc.

John R.

Ok, but what happens if the car stops, or changes speed while over the sensor? Wouldn't that throw the timing, and hence the data off? This is why commercial bar code scanners have a "sweep" or other mechanism so that the entire bar code is read at one time.

I have not ruled out the concept, just having too much fun tyring to get the RFID to work!

I personally don't care about the appearance. There are others out there (if my delusions of grandeur ever take off) that will object to all but the slightest (and even then some will object) change in the appearance of the cars, even on the underside. We call them "rivet counters" because they will notice that your box car has too many, or not enough, rivets along the door.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
John R.

8 + 8 = 10

Beau Schwabe

John R.,

If the car stops or changes speed to significantly loose data sync, then that data is dropped.
Hopefully you will have data from the other cars and write the necessary software that can
deduce where the "missing" car should be. Perhaps indicate a "warning" condition until the
MIA car can be found on another sensor in the track.


"This is why commercial bar code scanners have a "sweep" or other mechanism so that the entire bar code is read at one time."

That's why they work so well at the grocery store...

If the "Start Sync Bit" does not equal the "End Sync Bit" in my example above then the data can be considered invalid.


BTW) ....the train IS the "sweep" mechanism

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Beau Schwabe

IC Layout Engineer
Parallax, Inc.

Post Edited (Beau Schwabe (Parallax)) : 10/28/2005 3:05:58 PM GMT

John R.

The train being the sweep mechanism is a/the problem. The locations where you would typically want to have a "reader" are locations where speed and/or direction are changing, and/or the train will be stopping.

If I were to move from RFID to barcode, I believe I would end up going with some type of OEM barcode scanner. I haven't looked too hard yet, but I haven't found any cheap ones. I may have more luck grabbing one (scanner/engine) from a cheap wedge device. Even these are more expensive the RFID reader (assuming I can get it working).

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
John R.

8 + 8 = 10

Danny

Looking at the message above, the wedge devices are cheap, why not use a servo to move the cheap scanner? THen you can reliably read the bar codes no matter what the train is doing?

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
"Never create anything you can't control"
"The amount of intelligence on the planet is fixed... the population is growing"

John R.

A "cheap" non contact bar code scanner is about $100 and up. These include the mechanisms to do the scanning.

Using a servor to move a simple sensor like Beau is talking about might be an option if the RFID doesn't pan out. For that matter, and array of sensors might be an answer too. Interesting thoughts....

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
John R.

8 + 8 = 10

John R.

Beau;

As a piont of clarification on the stopping/reversiong, etc. A likely location for a scan point would be a siding where cars are dropped off and/or picked up. Tracking what cars went in and came out would be one of the main points. As this would be a situation where incomplete or corrupted data is likely, the very cars we're after tracking might be the ones that are dropped. (or a car gets scanned good going in, but "botched" (and the data dropped) on the way out, and then the system thinks the car is still in the siding).

The concept of seeing that car at another reader latter on may be workable. Actually the concept of using a combination of logic (like what cars should be in what order) and the scanning could also be applied to the RFID. In either case, it would be a good "sanity check" and could at least raise a flag for someone/something to check out.

Hopefully, all this will be mute, as the mighty god of electrons will smile on me this weekend, and all will be well with the world, and my system will be scanning RFID tags as the trains roll by at 200 MPH (scale speed). - I'd settle for even 100 MPH scale.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
John R.

8 + 8 = 10

Beau Schwabe

John R.

Another option that I had mentioned earlier... but would require each "train car" to siphon power from the track... in other words each car needs to be retro fitted with
metal pickup wheels. From there, each car could transmit an IR beacon of who he/she is. This would require a Stamp or SX (cheaper) in each car. This may or may
not be a practical approach.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Beau Schwabe

IC Layout Engineer
Parallax, Inc.