Hi,
we are using the ADXL335 in our product to measure angle of an arm.
We are using only the Z and Y axis.
We have the devices loaded onto a flex PCB in batches of 1000 by our Chinese manufacturer.
We are using 2.5V as the power supply. With 0g the output is expected to be 1.25V and vary up to +250mV and -250mV of this value when rotating through 360 degrees and incurring +1g to-1g.
However we are experiencing around 10% of the sensors with axis values outside of this range. Some a long way out. This causes our firmware to error and we are not able to accurately determine the angle.
We have studied the data sheet and it does provide max and min tolerance vales, but the faulty sensor axis values are still outside this range.
Can you comment on what may be the problem. Could this be a manufacturing issue. Could they be damaged? OR could they be second sourced non-original parts?
Thanks for you time,
Regards
Leon
Thank you for your post and for all of the information provided. I supposed that there could be reasonable question in all three areas of concern, which you provided. To start with, we are only able to provide support for products that have been procured through authorized distributors only. Are you in a position to verify that your subcontractor procured these products from an authorized distributor? All of the authorized channels are listed on this web site: https://www.analog.com/en/about-adi/corporate-information/sales-distribution.html
Assuming you are able to confirm this, can you share the -1g, 0g and +1g response measurements for all of your products, so we can study the distribution?
Thanks for the quick response.
You bet, Leon!
We look forward to hearing back from you!
Mark
NevadaMark,
can you please verify what I interpret from the data sheet.
Assuming a 3V supply, the typical 0g output will be 1.5V.
However the data sheet indicates that this could be from 1.8V to 1.2V and the chip would still be within manufacturing tolerance. Have I read this correct?
Also, If this is correct, does the +1g to -1g sensitivity range change? Typical sensitivity is +-300mV. But, for example, if the 0g offset is at the 1.8V level, is the +1g to -1g sensitivity altered?
Thanks
My understanding is that the bias errors (0g) are independent from the sensitivity errors. So, a part that has 1.8V output (0g), can have anywhere from +/-270mV to +/-330mV of variation, when rotated through the +/-1g gravity field. I am tagging a colleague of mine to be sure, as he has more direct experience with this product family. I hope that this helps.
jwang