Problem with Xbees which burnt due to datasheet error

I got 3 Xbees, two s1 and s2.
3 of them burned up. and this was what i have done:
1. connected 3v regulator to xbee suplly.
2. connected rx and tx to my mcu (at89c5131), Voh=5v Vol=0.

after a while they stop working and overload the regulator.
can it be becuase that mcu digital out is 5v? or can it be something else?

i got 1 xbee s2 left, i can do some tests, what to do?

BTW!
they got error on the data sheet, xbee Vih(max)=(0.8*vcc)-(0.18*vcc) and not 0.18*Vcc like they wrote! (they told me on support)

can i make digi send me new xbees because of this error?

Re: Xbee problams

Hi,
Very sad to hear your Xbee failed, but do not bother , it happens whenever you attempt new things, it infact adds more value to your final success and your sense of accomplishment.

As regards the connection of 5V logic output from microcontroller to the 3V powered Xbee is concerned, eventhough it is technically incorrect, if the micro port pin One level output is current limitted, as with most of the micros, I would suggest you to check for other reasons also for possible failue of the chip. By the way you also let me know how much is the current output at high level for the Micro pin.

If you can prove that the failure has come due to the mistake in the data sheet, you have all the reason to ask the manufacturer for a replacement of the failed unit.

Hoping that you would soon come out of your problem situation,

With regards,
Laktronics

Xbee problams

at the micro data sheet:
-60uA<Ioh<-10uA
1.6mA>Iol>100uA

i talk with digi support about the mistake at the datasheet and what i have done, they think also that it burns because of the high digital voltage of the micro. shuld i send them an email to ask for replacement?

Re: Xbee problams

Hi,
After one more reading of the Vih specification given by the support grup, it appears that the failure of the Xbee is mainly due to the 5V logic high applied to it. As per specification, the voltage is not to exeed 1.9V and even if you consider the error in data sheet, it is only more critical at 0.6V. Now you will also have problem with the output of Xbee to Micro input, as it will not satisfy the TTL specs. the micro needs, which means you will need suitable logic converters in between for interfacing.
Regards,
Laktronics

Added after 14 minutes:

Hi,
The specification you have quoted is with respect to the input of the micro when you apply a logic input to the micro. What you should find out is the output current at Voh, but never mind I will check it myself. As explained above, I think you have to accept the facts now and go for a replacement at your own cost only, but it does not prevent you from requesting ( not demanding) them for a replacement. Remember, it is always such little mistakes that adds up to one's experience.

Regards,
Laktronics

Xbee problams

it's my mistake, i decided not to ask for replacement.

to convert between interface i can use an PNP transistor, the digital output are with pull up resistors at the micro and at the xbee module:



it look like a good idea?

Re: Xbee problams

Hi,
It looks fine for me, I wish you did it the first time.
Regards,
Laktronics

Added after 55 minutes:

Hi,
But this 794 mv for 0 level is a bit high is n't it ? Try increasing 1K to some 3.3k and see how much voltage you get.
Regards,
Laktronics

Xbee problams

I realize that you want to use an emitter follower to keep from inverting the data, but its not a good idea. The emitter follower by design cannot drive a good logic low level. Regardless of the bias resistors, Vol will always be about one diode drop above ground.

I would suggest moving to a common emitter configuration with an NPN. You will need two of these in cascade to handle the logic inversion. This is assuming that the output of the micro is from a dedicated port like a UART. If you can produce one of the logic inversions inside the microcontroller via software, then you only need one external NPN in common emitter.

Philips, Fairchild and others make some dedicated logic converter ICs. These are designed to interface across logic families like this. In the final design you may want to consider these ICs as you can pack several signals into a signal chip with a much small board area.

Xbee problams

banjo, i dont think that it may be a problam, Vil on xbee is 0.2*vcc=0.6v, and the transistor Vce is less then 0.6 :O

but do you have an circuit exampe of your suggestion?

Xbee problams

I am no expert on Xbee, but I would guess that the 0.2*vcc is the Vil(max). They guarantee that the chip will work up to the max, but not beyond. Anything about 0.6V is may be read as a one, not a zero. Usually the zone between Vil(max) and Vih(min) is a dead zone where the logic level the chip will choose is undefined. You are not supposed to stay in this zone only rapidly transition through it.

It is not considered good design to operate right at the limit. For common emitter configurations, the governing parameter is Vbe and not Vce. The transistor cannot pull the emitter terminal lower than the base in common emitter. Since the base will be driven by the 5V microcontroller, its lower voltage limit could be around 0.2V. The emitter MUST be 1 diode drop above that. Even if you assume an 0.6V diode, this still puts the voltage at 0.8V. If the microcontroller pulls all the way to ground and the Xbee is forgiving, it may work for some devices at slow speeds.

For translation chips, look at:
http://www.standardics.nxp.com/support/techtips/simpleshifter/

Added after 2 minutes:

Sorry for the typo. Your proposed circuit is common collector. Vbe is governing parameter in common collector. I mis-typed the name of the configuration above, but the concept is the same.

Re: Xbee problams

Hi Ngel,
What Mr. Banjo has told are important points to be kept in mind when you deal with general logic circuits, many logic family will not tolerate high Vil levels and its noise immunity will reduce even if it works under lab conditions.
But in your specific case, Xbee quotes a figure of Vil = 0.35 *Vcc under electrical characteristics of their user manual which gives you adequate margin for your circuit and you have no reason to worry. In your case even a simple resistive divider with low value resistors might work. But for your future designs using multilogic devices do not underestimate the need for usage of standard logic converters.

Regards,
Laktronics

Added after 55 minutes:

Hi,
I also forgot to mention that you should keep an eye on possible other sources for failure of the Xbee. Make a thorough check up of all connections to Xbee. Check again the 3V power supply output for any transients on it. You can even connect a 3.6V zener across the 3V supply, with cathode connected to +3V. Check for any possible shorting of Xbee pins on the PCB. After making connections, just power up only one Xbee and feel its case temperature rise by touching it for a few minutes.
By the way, which signal is driving Xbee on the robot side?

Did a piar of them fail or only one end module failed, if a pair has failed, it almost points to the higher voltage input. Also you have told that for Xbee connection, you used a pull up on 5V side, how much was the resistor value? Intuitively I still feel that a current limitted drive should not have burnt Xbee.

Regards,
Laktronics

Re: Xbee problams

banjo thanks a lot for your explanation, i understand it and will keep working under the limit to keep the logic level right. the link is great and i will design right now a circuit that will convert and defend xbee.

laktronics, i will keep in mind to use zener for voltage limit. The xbee is connected on other developer board i build not on the robot and every one of the three xbee's connected to the same board, the pull up resistor is in the mcu (uart).

i will design a circuit and will post it here.

Added after 3 hours 45 minutes:

this is the circuit that i come up with.

Re: Xbee problams

Hi,
For XX07 you need a pull up to respective power supplies, 3V on Xbee side and 5V on Controller side.
Also, you have connected the LED to ON pin of Xbee with its anode to ground, needs change.
Regards,
Laktronics.

Added after 12 minutes:

Hi,
OK, pin ON is an output, so LED connection to ground is OK, but you need to reverse the direction of LED, you may need a series resistor also with it?
Regards,
Laktronics

Xbee problams

Oh just saw that error, well fixed it.
Hm. I will not need pull up resistors because the mcu and the xbee already have an pullup resistor so i dont need to add external pullup.
i dont think i will need a resistor to the led, xbee already have a current limit on that pin.