| ??? 10/27/10 16:09 Read: times |
#179370 - The original question was not answered Responding to: ???'s previous message |
esponding to: Hans Heinz's previous message
"If you use a 232A it is even worse." What is worse? using 10uf for the 232, which requires 1uF, you will have a factor of 10 error using 10uf for the 232A, which requires .1uF, you will have a factor of 100 error "What is worse?" is not an easy question to answer. The question what is worse is really asking what is the quantitative difference between the 232 and the 232A when changing the values of the capacitors. So, using your points 2 and 3: Per Westermark said:
2) With too small capacitors, the ripple will seriously hurt the communication. With too large capacitors, you get much longer charge time until the chip is able to perform the required work. And you may get into troubles with the transistors because when the chip is powered up, they have to work way harder to fill the capacitors.
3) Too big capacitors means they will also keep they charge for a longer time after you power off the device. This may be damaging to the chip depending on how the chip is designed - you have 10 times the energy that may potentially discharge over an unknown resistance in the chip. These two points do not tell us anything about the difference between the two chips. Now if you knew the empirical data that showed the specifics for 2 and 3 for the 232 and the specifics for 2 and 3 for the 232A then you may be able to come to a conclusion. This means that your two points are irrelevant in this situation. A simple fictional example, if using a 100uF cap on both 232 and 232A and they both fail, what is the cause of failure? If they both fail because they do not conform to standards in voltages, but are the same voltages, then neither is worse and they are the same. Maybe I am not making myself clear, but the empirical data will give a statistical quantitative result that will be the judge of "which one is worse". Stating that there is an error in the value of a factor of 10 or 100 makes little difference if the chips behave accordingly. In essence you cannot make a judgment without empirical data! Understanding what I am telling you is the reason for point 1: Responding to: Justin Fontes's previous message
1) You are the only one who somehow made the assumption that a factor 10 either had to be a factor ten more bit errors or a factor ten different baudrate or a factor ten different voltages or whatever you thought. Because his statements are incorrect in solving the question. |
