at Node A my discrepancy is -2.2802E-04 and at Node B discrepancy is 2.3418E-04. Since my currents do not = 0 at either node, what can I say about the size of the discrepacy? And about why the discrepancy with KCL might have occured.

 

Are you referring to a specific circuit? If so, can you post it. Otherwise the question you asked is unclear.

hgmjr

 

If you had used some approximating numerical method to solve the KCL equations, it might be the reason why you are not getting an exact zero.

 

at Node A my discrepancy is -2.2802E-04 and at Node B discrepancy is 2.3418E-04. Since my currents do not = 0 at either node, what can I say about the size of the discrepacy? And about why the discrepancy with KCL might have occured.

I know what we think, but what do you think? Are you ready to overturn KCL?

 

Probably for the same reason that if you enter 22/7*7 in most calculators, you'll get an answer of 21.99999999.

We had a teacher in tech school who was into microamps. He was soon replaced by a "real-world" teacher who told us calculations down to 2 decimal places was quite sufficient.

We trust our computers a way too much...

--Rich

 

And the sum of those two errors is 6.16 micro amps.

Does it really matter ... unless of course, if your test of interest concerns microamps.

228 microamps in a world of amps is zero. Would it really make a difference if you grabbed a circuit that produced 100 amps or a circuit that produced 100.000228 amps? I think not. Your physiological reaction wouldn't perceive the difference.

 

There is an academic thought regarding accuracy states that you should only define you answer in terms of the least accurate variable in your equations. Therefore, if you have two variables of 1.01 and 234.5, the least accurate variable is 234.5 (accurate to 1 decimal place). Therefore the product of these two terms is only accurate represented as 234.5 x 1 and not 1.01 - why because you can only define 235.5 to 1 dp, it could infact be 234.45 or conversely 234.54 having implications for your answer.

If you want to avoid inaccuracies (as much as possible) keep you answer in terms of stated fractions; 1/3A is more accurate than 0.333333333333333333333333333333333333333333333333A

Dave

 

Isn't the argument about the number of significant figures as opposed to decimal places. Especially when dealing with measured experimental data. It is questionable to combine two numbers with three significant figures each and express the result with 10 significant figures.

 

Isn't the argument about the number of significant figures as opposed to decimal places. Especially when dealing with measured experimental data. It is questionable to combine two numbers with three significant figures each and express the result with 10 significant figures.

Yes that is correct, particularly when the variables are not common, i.e. Amps x Ohms to get Volts. Since this is dealing with KCL my presumption in the above discussion is that the variables are common Amps plus Amps (although I have noticed I stated multiplied by mistake!) where, for example, 1.11A is stated more accurately than 1.1A.

The only way to state the sum of these with confidence of the degree of accuracy is 1.1 + 1.1 because to 1 dp (or in this example 2 sf) 1.1 is 1.1 and 1.11 is 1.1. To 2 dp (3 sf) 1.1 could be anything between 1.05 and 1.149.., where as to this degree of accuracy 1.11 is 1.11. The practical relevance to all of this is debatable of course!

Dave