Member Since: June 29, 2010

Country: United Kingdom

  • The schematic shows solder pads for pullups and tie-ing supplies together. What is the state of these pads on buying the device. Un-soldered makes sense - but the photos show big blobs of solder. Which is it?

  • Several questions regarding power: None of your images show which is Gnd vs +ve on the JST connector? What input voltage range can be used on the JST connector? The 3v3 pin on the side - is this an alternative way of powering it - or is 3v3 intended for output in which case how much additional current can be drawn? If I want to power it from a 5v board do I have to use the JST connector (extra alternative header pins would have been good!) or can I use 2 x 1N4001 diodes in series to drop the voltage and feed it in to the 3v3 pin on the side? Should I also use 100 ohm resistors in series on the clock and data lines if my mcu is working at 5v?

  • You are correct. But if your s/w lib activates pullups on the 5v mcu then you are stuffed. ie the internal AVR mcu 5v pullups of 50k plus the 4k7 on this board still give just over 3v3 and is probably ok. But if you add other breakout boards say at 5v with on board pullups then the SDA/SCL lines will creep towards 5v with the more you add. Until BMP085 goes bang. Worse - use a 5v mcu with a s/w lib that activates the internal mcu pullups and then remove the 3v3 supply to this sensor then SDA/SCL will pull up to 5v and it all goes bang. Happend to me! Connecting a 3v3 sensor like this to 5v mcu may work (depending on other i2c breakout boards you are using, what Vcc they use and whether they have onboard pullups). The wrong combination will kill ANY 3v3 sensors. Moral is - keep 3v3 sensors and 5v sensors on completely separate I2C buses. For the bus that matches your mcu Vcc then you can use its internal pullups - otherwise do NOT use the mcu pullups and insist the pullups are either added to one, or more, sensor boards or by you to the bus. ONE I2C BUS DOES NOT LIKE A MIXTURE OF DEVICES THAT EXPECT DIFFERENT VOLTAGES !!

  • Robert451,
    You are right. There is MAJOR confusion. I've got an older Razor and the magnetometer just doesn't work at all. Have asked SF about a hardware patch (see above) but heard nothing. You've tried the latest board and still have problems - which doesn't make me want to buy one after I've already bought one that doesn't work and they can't tell me how to fix it.
    As for code then I'm trying to create a plain vanilla AHRS/DCM system using WebbotLib (http://webbot.org.uk) that can be adapted to work with ANY gyro/accelerometer plus compass or magnetometer. Not released yet but if you want a .hex file to program the Razor with then give me a shout on webbot@webbot.org.uk
    Perhaps we can "get there" together since SF and their suggested AHRS code are totally out of sync.
    Perhaps SF should take lessons from the Mongoose board. The latter are clever enough to add an I2C header so that you can, at least, use other gyros/accel/magnetometers etc plus add any other I2C device.

  • Oops - now Robert451 has posted to say the magnetometer value is wrong. Did I speak too soon. The two magnetometers are the same other than the Y and Z values being reversed in the I2C registers.

  • Thanks for the quick response.
    It's re-assuring that this version of the board appears stable.
    Comparing the datasheets for the new HMC5883L and the old HMC5843 they both mention using low ESR caps, and removing any traces from under the chip. The only difference appears to be the addition of 0.1uF caps between Vdd and Gnd.
    Given that buying another board is not an option for me at the moment can SF give any official advice on how to 'fix' the older boards (ie adding the extra caps and/or replacing existing caps with low ESR ones). Or is it a more fundamental issue?

  • This is now the 3rd (released) revision of the Razor board and one would have thought that life would become easier. Alas not. The ability to download the correct firmware for the correct board from this site is a nightmare. Posts on each product forum would agree.
    Revision 1 of the board (http://www.sparkfun.com/products/9623) had a serious problem with the magnetometer (see http://forum.sparkfun.com/viewtopic.php?f=14&t=14756 and lots of other Google hits). This renders mine next to useless. After 2 weeks of playing the hardware only works once in a blue moon. Yes I've tried 6v unregulated to the battery header, +5v regulated to to the battery header, 3v3 to to the FTDI input etc etc. It just doesn't work! All the forum posts say is that I have to solder on a new SMD capacitor which I cannot do (why would I buy an SF breakout board if I can do my own SMD work?)
    1. What is SFs policy on returning items that do not work due to their own bad hardware design?
    2. Does such a policy apply to their worldwide dealers as well?
    3. Is the latest revision of the board free of such issues?
    4. Why would I want to risk spending another $125 to buy another version of their board when I have already spent $125 on a board that doesn't work - unless they guarantee that the answer to 3 above is 'Yes - it works 100% irrespective of power source'
    5. Isn't it rather telling that there are lots of other posts saying 'use the Mongoose firmware with the Razor' - when they could just say 'buy a Mongoose'.
    I don't expect to get any sensible reply from SF as they normally only reply to the posts that suit them. May be I'm better to save future money that I would spend on 'SF breakout boards' and spend it instead on kits to allow SMD ICs to plug into a breadboard.
    Rant over. Now waiting for the usual silence.....

  • Is there a reason why you still fail to support the definition of the custom characters? They are great for graphs, progress bars etc.

  • Not sure why Sparkfun went with the SMD board as the manufacturer (Hope) always used to sell a DIP/DIL version as well. So now I have to wait for a breakout board. Any time frame on that?

No public wish lists :(