| ??? 01/14/06 19:29 Read: times |
#107409 - RE : Answers IV Responding to: ???'s previous message |
Still way too high! Remember, output voltage of 741 is multiplied by factor 470/(27+25)=9 by the AD844! Again, only -0.177V is needed at C20 to provide the proper shifting. Changed it to provide -0.2 to 0 V. As the ADC must be read out very fast, the ADC must sit near the digital section. As also the AD844 must not be separated too far from ADC, it's best to have the analog and digital section on same board. The Digital board that I mean will place at the top of the Analog board. The cable that carries the AD844's output will be a shielded cable with less than 2 Cm length (soldered directly to the PCBs with no connectors). Analog Vcc and GND will be provided by a twisted pair with the same length of cable (Also with good decoupling both at the Analog PCB output and Digital PCB input). Would these solve the problem. The reason that I persist in having two different PCB is three reasons : * Since I have to get along with lower sample rates like 5 MSPS with an 16MIPS MCU managing the control, this would give the opportunity not to design a whole new board for the next version. * The final project will occupy less space , so it will be easier to use as a portable Data logger. But then you will need really proper filtering of the supply lines.
In any case, use separate ground planes! Sure, I'll use everything that I learn from you. Use the ferrites additionally to the resistors, means in series to them.
0.22µF tantals are not adequate! Use 10µF tantals in parallel to 100n X7R. Good news is that I found somewhere to buy SMD ferrite beads and SMD decoupling caps. I corrected every thing you mentioned. http://www.8052.com/users/far...pdated.GIF Thanks , Farshid |
