Wireless BMS?

[JRP3]

I've been intrigued by the idea of a wireless chip on each cell sending battery voltage to a receiver. I've been told that bluetooth probably won't work but RFID might. Does anyone know of an RFID chip that can monitor cell voltage, and maybe temperature? Is this possible? Or if not dynamic voltage then maybe a chip that sends a signal when a specific voltage is reached to shut off a charger.

 

[Tiberius]

Wireless data links in a BMS may sound like an outrageous idea, but it does actually make sense. Wireless is a neat way of getting around the differences in voltages and it can also be a way of getting round EMC problems. That last bit sounds counter intuitive, but all you have to is to run the wireless link above the frequency of the electrical interference.

First off, let's note that most BMSs already do use wireless links - in the form of optocouplers. Secondly, if anyone does want to seriously explore short range wireless links, my advice would be to engage a good RF consultant. PM me.

Let me explain a bit about RFID and Bluetooth and whether they are suitable. Both these are licence exempt forms of radio. In the US they fall under FCC Part 15; in Europe they are governed by various EN standards.

The common form of RFID is a way of putting a licence plate or ID number on objects in a way that can be read electronically. There is a low cost label on each object and a high cost reader. The reader generates a powerful radio signal and works out the ID number by what is reflected back. Now that alone might be a useful thing to do with expensive cells just for the purpose of tracking them through assembly and use. But if you want to do anything more, like check the voltage, you would probably need a custom chip, which would mean a start up cost of megabucks.

But it is a possibility. A custom RFID chip on each cell and a reader which would then gather the data and write instructions back would serve most of the functions of a BMS.

Bluetooth is a specification for a simple radio transceiver and a whole stack of operating software that allows it to communicate with any other Bluetooth device. Its geared up for transferring large amounts of data peer to peer at high speed. So its not really what is needed for a BMS for various reasons. But the idea of something like Bluetooth, but simpler and proprietary does make sense.

Again, a wireless device on each cell is a possibility. But it has to be compared against hard wiring, opto couplers, fibre optics, etc. Where a wireless link could really come into its own is communication from the BMS to the rest of the world.

HTH,

Nick

 

[JRP3]

So it sounds as if there probably is not a voltage reader RFID chip already out there, but the capability does exist? What companies make such things? Ideally this would be implemented at the factory on each cell and since it would make tracking and testing easier and faster would benefit the manufacturer as well as the consumer. However since that doesn't exist an after market solution would be useful. I'm not looking for the cell chip to take any action, just report data. I think I saw some RFID chips that can transmit temperature, what would be involved in adding voltage? Could it be done with a passive chip, or would you need active? I guess there is a third type that is passive until it's signaled then it becomes active?

 

[dak664]

RFIDs are usually powered by RF sent from the reader, an unnecessary complication for something already connected to a 10 amp-hour cell. A typical $2 microcontroller runs over a 2-5 volt range, has a 10 bit A/D converter and internal temperature sensor, and draws a microamp at idle. You could trigger it with a pushbutton, an RF pulse collected by a small antenna, or an infrared light pulse picked up by an LED, and bitbang an IR or RF response. One problem is selecting which cell to respond inside the battery pack; my solution was to have the first cell catch the trigger and report, then trigger the next cell up the voltage gradient, and so on. See http://endless-sphere.com/forums/viewtopic.php?f=14&t=13717

Full 802.15.4 transceivers could be added for an extra $5, giving ~30 meter range, that way each cell could have its own IP6 address and there would be no conflict. There are MCUs with integrated RF too, see http://www.atmel.com/dyn/products/product_card.asp?part_id=4692

 

[JRP3]

I think the active RFID chips are battery powered, so the cell could power them. The ATMEGA is interesting, but I expect once you put everything together to make it work the price is going to climb. I need wireless and cheap

 

[Tiberius]

JRP3,

Temp sensing is an obvious function for RFID and there are standard chips for that. Have a look at things like the MLX90129 from Melexis; it looks like that has a way of replacing the temp sensor with an external input, so it could read cell voltage.

There are all sorts of complexities, however, associated with the reading of these things. If anyone seriously wants to pursue it, pm me.

Nick

 

[JRP3]

There are all sorts of complexities, however, associated with the reading of these things. If anyone seriously wants to pursue it, pm me.

Nick

I did, check your PMs.

 

[Tiberius]

Hmm, one of my posts seems to have gone missing.

Anyway, John (JRP3) pm'd me, and since it was a general question, it might be better to answer it here. John was asking for more info about using wireless for cell level communications inside a BMS.

Wireless is always going to cost more, in development time and in manufacturing cost, than a wire and not be as good, except in special circumstances. There are also regulatory hurdles to overcome. What we have to ask is whether any such special circumstances apply here. Where wireless works is when one or more ends of the link is mobile, if running a wire is difficult, if the device in question has to communicate with many other devices, if the system is constantly being reconfigured, etc, etc.

Generally none of these things apply to a cell inside a pack. Moreover, the cells have to have power wires connected, so wiring up the BMS signals isn't a big deal. What might make it worth looking at wireless is if very expensive cells have to be tracked, if external devices need to check on the cells, if there is a lot of swapping in and out of cells, etc, etc. But even then, wireless cell connectivity has to be compared with wiring the cells to master unit and using wireless from there.

I wouldn't rule it out, but my guess is its not worth doing unless you identify a need for an external device to talk to an individual high value cell. That could happen as part of systems for EV infrastructure, especially if batteries are leased, but that is getting way beyond normal BMS functions.

Another thing to remember is that a wireless link is never 100% reliable. More than anything else it is prone to work on the bench and fail in the field. A lot of my consultancy work is with people who have a wireless system that works 90% of the time, when they really needed 99.9%.

To paraphrase something often quoted on ES. You can have any two out of this three:

• Low manufacturing cost
  Low design cost
  It working properly.

There is some work going on towards using frequencies around 50 to 100 GHz for very short range data links inside electronic units, perhaps even from one side to the other of an IC. That might be neat for cell level communications, but I don't think its ready yet.

HTH,

Nick

 

[JRP3]

Good summation. Probably would only be cost effective if implemented at the factory level, where tracking and monitoring the cells would be beneficial, and then since it's already in place it provides added value to the product in the field. If it works.

 

[Tiberius]

I should add some background here. A number of people on ES know that my day job is running a wireless design and consultancy business, which puts me in a slightly awkward position in this thread.

I'm happy to offer general comments, give overviews, explanations and help correct misapprehensions. I'm even happy to explain how RFID works, what it can and can't do, and talk about some of the regulatory issues involved in marketing wireless products. I don't mind sharing public domain knowledge and teaching people the basics.

What I was thinking of when I said pm me if you are serious, was that if anyone needed professional help on a commercial or industrial project, then we should obviously take the discussion private.

Nick