I remember this from before, and would like to explore it a little further (for other reasons).· Basically, I'm thinking that this could also most work like a R/Q RFID, without the RF.
The track has power that can easily be siphoned.· DCC has power on the rails at all time, along with a signal for control of the trains.· Getting power is not a problem.· How fast can the IR transmitter/receiver move data?· (I know, I could look this up, but I'm an american!)
In concept I'm thinking about a system that would have the ID of the car somehow hard coded into the onboard component of the system, along with a way of dynamically recording point of origin and destination.· Then as the car went past various sensing points, if it was supposed to be dropped before the next station, that could be displayed on an LCD (or computer screen).· For this, you'd have to have provisions for not only transmitting the ID, but also a way to send information to the SX (or STAMP) and receiving and storing the data.· We'd also ideally do this as the train is moving.· Is this possible?
Even with an SX chip, the cost may not be affordable, but I like the concept.
Beau Schwabe (Parallax) said...
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.
That's ok, but I think you need to focus on some of the basics first.
Going back to the passive RFID concept, there was mention about an active
RFID solution. This spawned the idea of an IR beacon solution which effectively
would be an active IR solution.
After thinking about this further, there is a way to create a "dumb" IR transmitter
with discrete components. There are two frequencies that need to be supplied.
One is the baud frequency, and the other is the IR carrier frequency. Both can be
generated with a 556 ( a dual 555 ) . In the schematic below, it works out that the
baud rate is the same as the input frequency in Hertz.
Another thing.... 200mph at 1/160 scale would be 1.25mph as you or I would see it?
1.25miles = 6600 feet (5280feet X 1.25 miles = 6600feet)
6600 feet / 3600 seconds ( in one hour) = 1.83 feet per second
Even at 300 Baud..... 33.3 mS per BYTE .... You’re going to get about 30 samples
per second.
1.83feet = 22 inches
Or looking at it another way 1.36 samples per inch (30 samples per second / 22 inches
per second = 1.3636...) at 300 Baud at 200 mph.
so if you want at least 5 samples at that speed, your IR sensor needs to have a range
of at least 3.7 inches. (5 samples divided by 1.36 samples per inch = 3.6666 inches)
Even in a worst case IR sensor scenario 4 inches is nothing, this should be completely
doable.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔ Beau Schwabe
IC Layout Engineer
Parallax, Inc.
From my perspective, the design specs aren't changing, but the discussion has also been looking at alternatives (and potential enhancements) should the original idea end up floundering. I will take blame for alternately discouraging and encouraging discussion of alternative possibilities.
As clarification for all (possibly including my self):
Immediate Project:
The current project is focused on using the Parallax RFID reader, and compatible tags from alternate sources (due to the relativey "huge" size of the tags from Parallax).
The application is to affix a tag to model railroad cars (one to each) and read the tags as the train passes. This data will ultimately be fed to a PC. The specifics of getting the data to the PC, and what to do with it once it gets there, are a secondary project. The reader antenna will be hidden underneath the track in the preferred design configuration. I have "delusions of grandeur" of a commercial application to be marketed in the hobby community. This may, or may not prove viable. On one hand, this is applied technolgoy for technology's sake. On the other, this (RFID) is how rail cars are currently tracked in the "real world" (not realitiy TV, but the place where real people live and work).
The current "problem" is getting a tag and antenna combination to work in a Model Railroad environment. I am starting with N Scale/Gauge = 1/160 scale. I plan on then working up to larger scales, or in a "worse case" scenario, skipping N and working to larger scales. My personal railroad is On30 (1/48 actual, narrow gauge). I have had moderate success in attaching a 13mm "capsule" to the bottom of a pair of test cars (N Gauge), and have the antenna below the track ties (but above the cork roadbed). This only works at slow (very slow) train speeds, but has demonstrated a "proof of concept".
Planned action:
External to a rail car, map the response of the antenna to various tags in terms of distance and orientation. The antenna/tag combination has very different results in different areas of the antenna, and with different orientation of the tags.
Explore using "RF Reflective" material (sheet metal, screen, etc.) to shape the RF Field of both the antenna and tags to enhance system performance.
Prototype "best case" combination(s) from above testing and experimentation.
Research reader antenna requirements and any possible design alternatives via the WWW and the chip OEMs site. Possibly attempt to fabricate an alternative antenna.
Evaluat status.
Alternative Projects
The above project has been fermenting in my mind for several years. (Since doing some RFID testing at work. I'm an IT manager, and at my previous job, we explored and prototyped using R/W tags for inventory tracking and automating transactions.) Until the recent discovery of the Phidgets RFID reader, and more significantly, that leading to the discovery of the Parallax reader and STAMP and SX kits, etc. it has laid dormant. There is a LOT of internal momentum to keep this project alive, to the point of probably ignoring any expert advice saying it can't be done, and continuing to plow ahead. (I am not implying that anyone has said that, but if the world's leading expert walked up and said "no way", I'd probably ask her/him to either shut up and provide constructive comments, or leave...)
On the other hand, there is a realistic side of me that says "don't ignore other possibilities", and some of this thread (including the bi-directional IR) has been looking at those ideas. Even when (if) I get the RFID working, these other ideas may get some actual development. Model Railroading for me is a way to explore other things I have an interest in, and bring them together in an "applied" way. This includes mineature animations (working cranes, etc.) and the application of techology (sometimes purely for the sake of technology). My railroad stuff generally doesn't look too prety, but I have fun, and that's what counts.
Summary
So this thread may meander, and I have enjoyed all the alternative discussions, even those I may have seemed to "bash down". Those ideas may become significant, and revisited, pending the relative success/failure of the immediate project.
Keep the ideas coming, both those specific to the RFID, as well as all those wonderful alternatives. Kind of like a brain storming session. There are wonderfly two things going on, the specific project, and discussion of alternatives. It has been great seeing all the activity.
I hope this gives all a better understanding of all my ramblings.
In concept I'm thinking about a system that would have the ID of the car somehow hard coded into the onboard component of the system, along with a way of dynamically recording point of origin and destination. Then as the car went past various sensing points, if it was supposed to be dropped before the next station, that could be displayed on an LCD (or computer screen). For this, you'd have to have provisions for not only transmitting the ID, but also a way to send information to the SX (or STAMP) and receiving and storing the data. We'd also ideally do this as the train is moving. Is this possible?
I guess this is what threw me off....up to this point we have been talking basically about a one-way form of communication.
Adding the element of bi-direction complicates things a bit, not that it is impossible, but more difficult. I apologize if I sounded
a bit tempered, I did not mean for it to come out that way.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔ Beau Schwabe
IC Layout Engineer
Parallax, Inc.
In another post somebody was looking to build an RFID controlled pet door. I was going to suggest using an iButton for the ID, but they are a contact device, not RF. I was thinking about whether they could be used here, and here is a theoretical possibility, and let's use HO scale for a common denominator:
1. Put an iButton in each car
2. Put all metal Kaydee or similar couplers on each car
3. Wire the iButton to the coupler, so that when cars are coupled, there is a signal path between the buttons through the couplers
4. Wire the whole train this way, including locomotive
5. Add some wiring/circuitry in the locomotive to interface the iButton chain to the track, so the wheels are the signal connection
6. Interface iButton readers to the track in some manner that achieves the desired result
If this is do-able, I think you could put a Stamp/SX in the locomotive, connected to the iButton circuit. You should then be able to poll for the ID of every car attached. Dropping a car drops the ID for the circuit, picking up a car adds it to the circuit. I think the biggest challenge would be to get all of the signals properly connected, including the reader in the track. But since the iButton simply requires contact, you should only need to complete the circuit. Also, Maxim has a name for the concept: a 1-wire mirco-lan. Here is a blurb from their website:
The DS9092R and DS9092RG are 1-Wire Ports designed for communicating to 1-Wire chips or iButtons within an embedded system. These 1-Wire ports are simply empty iButton 'cans' with data and ground leads inside to allow easy wiring to a microprocessor or 1-Wire micro-LAN. If a product contains 1-Wire chips and/or iButtons, these ports allow you to communicate with all the 1-Wire devices simply by touching the port with one of the 1-Wire probes described above. Use the RG version for touch-and-hold applications.
So based on that description you could probably wire the port to the track, and connect it to a probe under the layout, and interface the probe to your PC.
Comments
The track has power that can easily be siphoned.· DCC has power on the rails at all time, along with a signal for control of the trains.· Getting power is not a problem.· How fast can the IR transmitter/receiver move data?· (I know, I could look this up, but I'm an american!)
In concept I'm thinking about a system that would have the ID of the car somehow hard coded into the onboard component of the system, along with a way of dynamically recording point of origin and destination.· Then as the car went past various sensing points, if it was supposed to be dropped before the next station, that could be displayed on an LCD (or computer screen).· For this, you'd have to have provisions for not only transmitting the ID, but also a way to send information to the SX (or STAMP) and receiving and storing the data.· We'd also ideally do this as the train is moving.· Is this possible?
Even with an SX chip, the cost may not be affordable, but I like the concept.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
John R.
8 + 8 = 10
Your design parameters keep changing on us...
That's ok, but I think you need to focus on some of the basics first.
Going back to the passive RFID concept, there was mention about an active
RFID solution. This spawned the idea of an IR beacon solution which effectively
would be an active IR solution.
After thinking about this further, there is a way to create a "dumb" IR transmitter
with discrete components. There are two frequencies that need to be supplied.
One is the baud frequency, and the other is the IR carrier frequency. Both can be
generated with a 556 ( a dual 555 ) . In the schematic below, it works out that the
baud rate is the same as the input frequency in Hertz.
For example: 300 Baud = 300Hz : 1200 Baud = 1200Hz : 9600 Baud = 9600Hz
Another thing.... 200mph at 1/160 scale would be 1.25mph as you or I would see it?
1.25miles = 6600 feet (5280feet X 1.25 miles = 6600feet)
6600 feet / 3600 seconds ( in one hour) = 1.83 feet per second
Even at 300 Baud..... 33.3 mS per BYTE .... You’re going to get about 30 samples
per second.
1.83feet = 22 inches
Or looking at it another way 1.36 samples per inch (30 samples per second / 22 inches
per second = 1.3636...) at 300 Baud at 200 mph.
so if you want at least 5 samples at that speed, your IR sensor needs to have a range
of at least 3.7 inches. (5 samples divided by 1.36 samples per inch = 3.6666 inches)
Even in a worst case IR sensor scenario 4 inches is nothing, this should be completely
doable.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Beau Schwabe
IC Layout Engineer
Parallax, Inc.
From my perspective, the design specs aren't changing, but the discussion has also been looking at alternatives (and potential enhancements) should the original idea end up floundering. I will take blame for alternately discouraging and encouraging discussion of alternative possibilities.
As clarification for all (possibly including my self):
Immediate Project:
The current project is focused on using the Parallax RFID reader, and compatible tags from alternate sources (due to the relativey "huge" size of the tags from Parallax).
The application is to affix a tag to model railroad cars (one to each) and read the tags as the train passes. This data will ultimately be fed to a PC. The specifics of getting the data to the PC, and what to do with it once it gets there, are a secondary project. The reader antenna will be hidden underneath the track in the preferred design configuration. I have "delusions of grandeur" of a commercial application to be marketed in the hobby community. This may, or may not prove viable. On one hand, this is applied technolgoy for technology's sake. On the other, this (RFID) is how rail cars are currently tracked in the "real world" (not realitiy TV, but the place where real people live and work).
The current "problem" is getting a tag and antenna combination to work in a Model Railroad environment. I am starting with N Scale/Gauge = 1/160 scale. I plan on then working up to larger scales, or in a "worse case" scenario, skipping N and working to larger scales. My personal railroad is On30 (1/48 actual, narrow gauge). I have had moderate success in attaching a 13mm "capsule" to the bottom of a pair of test cars (N Gauge), and have the antenna below the track ties (but above the cork roadbed). This only works at slow (very slow) train speeds, but has demonstrated a "proof of concept".
Planned action:
External to a rail car, map the response of the antenna to various tags in terms of distance and orientation. The antenna/tag combination has very different results in different areas of the antenna, and with different orientation of the tags.
Explore using "RF Reflective" material (sheet metal, screen, etc.) to shape the RF Field of both the antenna and tags to enhance system performance.
Prototype "best case" combination(s) from above testing and experimentation.
Research reader antenna requirements and any possible design alternatives via the WWW and the chip OEMs site. Possibly attempt to fabricate an alternative antenna.
Evaluat status.
Alternative Projects
The above project has been fermenting in my mind for several years. (Since doing some RFID testing at work. I'm an IT manager, and at my previous job, we explored and prototyped using R/W tags for inventory tracking and automating transactions.) Until the recent discovery of the Phidgets RFID reader, and more significantly, that leading to the discovery of the Parallax reader and STAMP and SX kits, etc. it has laid dormant. There is a LOT of internal momentum to keep this project alive, to the point of probably ignoring any expert advice saying it can't be done, and continuing to plow ahead. (I am not implying that anyone has said that, but if the world's leading expert walked up and said "no way", I'd probably ask her/him to either shut up and provide constructive comments, or leave...)
On the other hand, there is a realistic side of me that says "don't ignore other possibilities", and some of this thread (including the bi-directional IR) has been looking at those ideas. Even when (if) I get the RFID working, these other ideas may get some actual development. Model Railroading for me is a way to explore other things I have an interest in, and bring them together in an "applied" way. This includes mineature animations (working cranes, etc.) and the application of techology (sometimes purely for the sake of technology). My railroad stuff generally doesn't look too prety, but I have fun, and that's what counts.
Summary
So this thread may meander, and I have enjoyed all the alternative discussions, even those I may have seemed to "bash down". Those ideas may become significant, and revisited, pending the relative success/failure of the immediate project.
Keep the ideas coming, both those specific to the RFID, as well as all those wonderful alternatives. Kind of like a brain storming session. There are wonderfly two things going on, the specific project, and discussion of alternatives. It has been great seeing all the activity.
I hope this gives all a better understanding of all my ramblings.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
John R.
8 + 8 = 10
I guess this is what threw me off....up to this point we have been talking basically about a one-way form of communication.
Adding the element of bi-direction complicates things a bit, not that it is impossible, but more difficult. I apologize if I sounded
a bit tempered, I did not mean for it to come out that way.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Beau Schwabe
IC Layout Engineer
Parallax, Inc.
You didn't sound tempered at all, and the emoticon made that even clearer.
Given the length and various directions of the thread, I thought clarification was in order.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
John R.
8 + 8 = 10
1. Put an iButton in each car
2. Put all metal Kaydee or similar couplers on each car
3. Wire the iButton to the coupler, so that when cars are coupled, there is a signal path between the buttons through the couplers
4. Wire the whole train this way, including locomotive
5. Add some wiring/circuitry in the locomotive to interface the iButton chain to the track, so the wheels are the signal connection
6. Interface iButton readers to the track in some manner that achieves the desired result
If this is do-able, I think you could put a Stamp/SX in the locomotive, connected to the iButton circuit. You should then be able to poll for the ID of every car attached. Dropping a car drops the ID for the circuit, picking up a car adds it to the circuit. I think the biggest challenge would be to get all of the signals properly connected, including the reader in the track. But since the iButton simply requires contact, you should only need to complete the circuit. Also, Maxim has a name for the concept: a 1-wire mirco-lan. Here is a blurb from their website:
The DS9092R and DS9092RG are 1-Wire Ports designed for communicating to 1-Wire chips or iButtons within an embedded system. These 1-Wire ports are simply empty iButton 'cans' with data and ground leads inside to allow easy wiring to a microprocessor or 1-Wire micro-LAN. If a product contains 1-Wire chips and/or iButtons, these ports allow you to communicate with all the 1-Wire devices simply by touching the port with one of the 1-Wire probes described above. Use the RG version for touch-and-hold applications.
So based on that description you could probably wire the port to the track, and connect it to a probe under the layout, and interface the probe to your PC.