| ??? 10/22/07 18:14 Read: times |
#146016 - Hello Erik Responding to: ???'s previous message |
Erik Malund said:
take as an example an 8 bit ADC where 1V raeds 047h and 1,01v reads 048h. what do you expect if the input is 1.005V/ a tolerance/stability of a picovolt??? Hmmm.... 1.005V > 0.005 Picovolts?? mVolts surely? Erik Malund said:
have you read http://www.8052.com/faqs.phtml?FAQ=128848 I guarantee you that even if the last bit stopped "jumping around" your precision would be nothing like you think. Yes, thats very true Erik, but I'm getting there with minimal costs. First job is to stabilise my reading and once achieved I will work on the accuracy. Erik Malund said:
scales that have an actual precision (do not read 'resolution' DO read precision) cost in the order of $10k, where did you get the illusion that you can do better for a few bucks? A few bucks you say!, thats too high, the TC500A costs around $3.00 (I can probably get it even cheaper from China) but it has an add on adapter, the TC520A, the latter simplyfies the driver software overhead but restricts the TC500A to 17 bits, I'm not using TC520A but instead drive the TC500A directly from the MCU, pushing 20 bits resolution out of it. At first, I was concerned about ADC linearity, but having tested it, it's as linear as an arrow even at 20 bits. Semi manufactures always build in contingency just to cover themselves. Lets face it, both Fluke and Agilent are at the forefront of measurement technology? The thing is one can download service manuals complete from Agilent/fluke totally free. Reading and studying these manuals gives insight into how these instruments work which is what I have been doing for some time. Cheers for now Jason |
