Does the Parallax USB to xBee adaptor support using xBee Pro - 250ma > 100ma

Zeph

I am building a sensor network with the xBee and xBee Pro (series 1).

I bought the Parallax USB xBee adaptor, not just to program the xBee, but to use as my interface from monitoring computer (Windows laptop) to the xBee nodes, because it specifies that it supports the xBee Pro; there were cheaper versions on eBay but I trusted Parallax. I'm trying to find a problem.

Now I am digging deeper to find that the adaptor uses the FAN2500S33X regulator which provides 100ma at 3.3v, while the xBee Pro specifies a typical power supply draw of 250ma when transmitting. It would seem that the Parallax adaptor will not support actually *using* the xBee Pro to transmit, unless I'm missing something.

Is this accurate? (It would be nice if Parallax mentioned this limitation - "supports xBee Pro for programming and receive only")

Are there any workarounds, or was this a wasted purchase? (I now see that a competitor has a 500ma 3.3v regulator)?

Thanks for any tips.

Edit. While the doc for this product #32400 (http://www.parallax.com/Portals/0/Downloads/docs/prod/rf/32400-XBeeUSBAdapter-v1.0.pdf) has no warning (you have to check the schematic for the regulator and search for the part), it turns out that Parallax is aware of this issue because the related non-USB adaptor (#32401) does have a warning in the doc file (http://www.parallax.com/Portals/0/Downloads/docs/prod/rf/32401-XBee5V3.3VAdapter-v1.2.pdf).

In the documents for product 32401 it indicates that the regulator can handle brief transmissions. Can I assume this is true for product 32400 as well?

Franklin

You can use a large cap as a temporary power source if you keep your transmits moderatly short.

Zeph

Franklin wrote: »

You can use a large cap as a temporary power source if you keep your transmits moderatly short.

Something like 100uF on the output side of the FAN2500S33X? (Wired to the xBee socket or 0.1" headers probably?)

Mike Green

A large capacitor across the regulator output won't help you that much. Even a 1000uF capacitor will only give you a couple of milliseconds of power at 250mA. If the transmissions are short and infrequent, the thermal lag of the regulator will make the most difference. The regulator is thermally limited, not current limited. With a 5V input, the regulator will have to dissipate (5V - 3.3V) x 0.25A or just under 1/2W. That's a lot for a little package on a little PCB. On the other hand, if it's transmitting for a few milliseconds maybe once every few seconds, the average power dissipation will be low enough so the regulator won't shut down. You'll have to figure out the duty cycle for your application and do some testing to see what the temperature rise is.

Keep in mind that, if you cut things too closely in terms of heat, a small rise in USB power supply voltage or ambient temperature might shut things down. Also remember that USB ports only supply a limited amount of power. It's 500mA, but only after the computer enumerates the device and only if the device tells the computer that it's a high power device.

Phil Pilgrim (PhiPi)

Rev B of #32400 has a much more robust regulator. I've had no issues using it with the XBee Pro. That's the version that Parallax is shipping now, BTW.



-Phil

Zeph

Thanks Mike for the tip about what the actual limitation is. In terms of dong some estimates, how would I estimate how long I could draw 250ma (dissipating about 1/2 W)? A useful result would be something like "keep average dissipation under 200mw within any 100 msec period", in which case I could use the RX, TX and Idle currents to estimate a duty cycle, and design around that. I don't know anything about doing dynamic heat estimations (like what the averaging period would be).

Also - how would I estimate transmit time? Is the on-air packet similar to the API frame plus some number of bytes of sync prefix? It would be useful if I could say "If I send a 20 byte API frame, it will result in 24 synchronous 8 bit bytes on air at 250,000 bps so TX power will be drawn for about 770 us". Anybody know how to make such an estimate?

And Phil - thanks for letting me know I must have bought some of the last Rev A, sigh. I guess it's not likely I can find anybody who wants to swap...

Mike Green

Let's say that the transmit speed of the xBee is roughly 100,000 bps. A 100 byte packet might take 800 bits rounded up to 1,000 bits to include a header, checksum, etc. That's maybe 100 packets per second, so a packet might take 10ms. Write a program that sends 100 bytes of data from one xBeePro to another once a second for a minute, then 10 times a second for a minute, then 100 times a second for a minute. Get a thermometer of some kind that's able to read the temperature near the regulator. If there's not enough room, make a probe with aluminum foil rolled into a short thin rod covered with a thin plastic tape (for electrical insulation) and coupled thermally (and mechanically) to the thermometer. Twisting some of the foil around the thermometer might do. Run the program and see what happens to the temperature. A minute should be enough time for the temperature to stabilize at each step.

You might be able to use the mounting hole next to the regulator as a measuring point

Zeph

Zeph wrote: »

Also - how would I estimate transmit time? Is the on-air packet similar to the API frame plus some number of bytes of sync prefix? It would be useful if I could say "If I send a 20 byte API frame, it will result in 24 synchronous 8 bit bytes on air at 250,000 bps so TX power will be drawn for about 770 us". Anybody know how to make such an estimate?

Partially answering my own question about time. The xBee (2.4 GHz 802.15.4 version, ie: series 1) claims to have a 250,000 bps data rate; that's presumably the on-air bit rate since the async serial interface is configurable and generally lower; they probably have a packed synchronous packet format (8 bits on air per byte without start and stop bits) with packet headers etc. It's likely that the on air packet has a similar format to the API frame, plus a few bytes for sync. So I can probably estimate the TX time well enough from that.

Mike - I gather your suggestion is to empirically measure temp rise in the regulator with various loads, rather than attempting to calculate it. That I understand. (I would like to someday understand a bit more about non-static thermal calculations, but it can wait). Any ideas on how to translate the externally measured temperature to junction temps for the SMD package of the regulator?

Thanks.

Thanks

Chris Savage

This thread is old and I am assuming it is closed/solved?

I just wanted to add in regards to the regulator that the 100mA rating is the nominal current output used in all the testing. Ideally you would not have more than 100mA continuous. The FAN2500 is rated at 300mA max output, not under absolute maximum specs, this is just the upper normal limit. So under normal use with the USB XBee Adapter there should not be an issue. As has been pointed out the regulator was upgraded in the Rev B. and higher boards.