Warehouse RFID tags

[shantanu@india]

Hi,
I am working on a warehouse inventory management project where about 5000 products are stored on racks.The size of the warehouse maybe 100 X 100 meters. They need a system to know the exact stock position by means of reading the RFID tags attached to the products.
A search of RFID yeilded the information that most RFID tags are passive in nature(13 MHz) & they work on harnessing energy from radio wave principle. The working distance of the RFID reader from the tag cannot be more than 100 mm.(I may be wrong...someone might have clearer concept!!)
I was thinking along different lines...suppose I combine a 433MHz low cost RF module with a basic PIC like 12F675 & power it from a button cell.It would be always on sleep mode & would wake up on a RF identification call & the go back to sleep again. Whenever a product enters the warehouse a tag is programmed with an ID code and fixed to the product & whenever the product leaves the warehouse that tag is erased.
Would this system work or can there be better solutions?
Thanks & Happy New Year everyone.

[top204]

The problem with the standard receivers is that they still draw quite a lot of current when in receive mode. So the battery would not last long.

The RFID tags are extremely inexpensive, and the reading of them is very straightforward. They come in 125KHz and 13MHz types, where the higher the frequency, the shorter the range they have (generally).

[JonW]

Stock position as in location or number of products in the warehouse?

Passive RFID can work over a much larger distance depending on the technology in the tag.  Some of the longer distance tags work at 2.45GHz or 5GHz plus and can work over meters.  Inventory management can be passive as long as the products are correctly registered in and out of the facility.

If you want an instantaneous stock level using RF then its much better to use a single master receiver and tag the product with a transmitter.  This way you really reduce the power to uA, wake randomly and transmit avoiding collisions with a decent protocol.   

Lowest cost would be using something like the CMT2119 RFIC with a much lower cost MCU like NY08A50X series of OTP MCU from Nyquest or similar, print a simple antenna on FR4.  The MCU were around $0.02 before the Semi market went pop and the RFIC less than $0.10, I have designed these MCU's in to alot of products, they have a flash version but cost is higher. You can easily give each tag a unique ID and allocate to a product on entry and deactivate/reactivate using non contact methods.  Some have ADC so you can report battery status, temperature or other info if needed.

[shantanu@india]

Thanks John for your ideas. Its really stock position for the time being though physical location might come later.
I'll research on the devices that you mention.

[normnet]

The RFID supplier Daily may be of some interest.

[shantanu@india]

Thanks Norm

[shantanu@india]

What about rfPIC's?
Has anyone usedthem?

[shantanu@india]

Another possible option using true IOT is to use  ESP8266 modules powered from 3V button cells. NodeMCU or equivalent modules are very cheap and are equipped with WiFi. As usual, we can write code to wake up the modules @ 1 hour and transmit an identifier string. The modules would have unique IP addresses and MAC addresses so that even if there are thousands of modules inside a warehouse, there cannot be any chance of collision since web server client technology supported by the solid TCP/IP framework.

[JonW]

Still not going to be as low cost as the original solution with a basic MCU and transmitter.  Issue you will likely face is getting enough devices in the short term and then lead times, I would start by seeing what's available and build on it.  A reel or two of low cost MCU/TX RFIC wont break the bank and keeping stock will allow you to manufacture more tags at short notice. You can start the design using a 10F200 or similar and then port to a cheaper MCU with a cost down exercise once the product is mature.

ISM bands 433/868 MHz are extremely cost effective for short range data transfer and very easy to interface to. Receivers are also cheap, easily obtained, reliable and will have good coverage.  ESP and 2.45G will likely have poor coverage and be susceptible to interference.  You may need to do a site survey to see what kind of system you need to implement as its unlikely you will get line of sight comms and tags maybe covered. 

Other option is to go for LoRa but again, cost availability of devices will be an issue.  Many years ago Hitachi developed a RFID chip called Mu, Hong Kong airport used it to automate and track luggage but again these will require readers, writers and MOQ so you are back at Passive RFID again. 

[shantanu@india]

Thanks for your detailed advice John. I shall make a few tags according to your design plan.

[JonW]

No problem.  If you need a printed antenna on FR4 let me know and I can design one for you, just need your substrate thickness, material and frequency. 0.6mm Fr4 would be best, 50R trace thickness approx 1.05mm  Alternative is to use an SMT chip antenna to get the pcb as small as possible and this will work first time and you don't have to worry about optimising the antenna on a VNA.  Easily be able to fit the whole design and antenna a small coin cell diameter with a chip antenna

Jon

[Trikkitt]

Another possible option using true IOT is to use  ESP8266 modules powered from 3V button cells. NodeMCU or equivalent modules are very cheap and are equipped with WiFi. As usual, we can write code to wake up the modules @ 1 hour and transmit an identifier string. The modules would have unique IP addresses and MAC addresses so that even if there are thousands of modules inside a warehouse, there cannot be any chance of collision since web server client technology supported by the solid TCP/IP framework.

Just one word of warning - never assume unique MAC addresses amongst the cheap ESP8266 modules.  The suppliers aren't registered to allocated MAC addresses, and so they just make them up, often repeating or overlapping blocks.  I ordered about 40 modules back in 2018 and at least 3 had duplicate MAC addresses.  It is possible to re-flash them I believe, but I didn't bother as I had spares.  However it is something to look out for, especially with cheap stuff that is unbranded or has an unknown origin despite being branded!