Why use 16-bit analog cards?

------------------------------------------------------------------------------------------
Split from Analog scaling .

Gustavo Banchetti

This post brings to mind an intimate question that I have always asked myself and to which I have never really given an answer. How does Siemens justify the sense of using 16-bit analog cards, only to actually lose definition in the user program while still limiting the bandwidth of useful value to 14 bits (and counting)?

Now I know the moderators will go and split my post making you lose comprehension as well, but I would be curious to know a serious reason. Not hypothesis.

I have only a partial answer for you. You already know what I am putting down here. I am the wrong person to start your argument with since I cannot speak to Siemens' motivation for making such design choices

The specification does say that one bit is used for the sign. So that means for a single sided measurement (0 - 10V / 4 - 20mA), there is only a "15" bit resolution. 

Assume the measurement is 0 to 10V. The actual measurable voltage value is higher than 10V. The following table comes from the S7-1200's system manual. Observe the over ranges.

This is in fact what I know and what I suppose. Ok for the sign bit, where the 16 bit card are intended in general for bipolar signals. But not ok for the 'cutting' of the significat 15 bit's. Generally the Siemens A/D conversion is intented from 0 % to 110 % of the allowable analog signal, resulting in a number from 0 up to 30000 (and counting), where in fact we use the 0 .. 27648 bandwidth as 0% .. 100%

I can accept that in industrial use this kind of conversion is generally acceptable. I can accept that in the same numerical tag there are also the main diagnostics informations. But I can't accept that the card are selled as 16 bits, where in fact the resulting data is treated as 14 bits.

I much prefer the solutions from other manufacturers that provide diagnostics on bits outside the conversion word, but do not cut the data, perhaps forcing the programmer to write the filters and various offset corrections himself, but without losing anything in the definition of the 16 bits, which in some circumstances remain welcome.

Gustavo Banchetti

This is in fact what I know and what I suppose. Ok for the sign bit, where the 16 bit card are intended in general for bipolar signals. But not ok for the 'cutting' of the significat 15 bit's. Generally the Siemens A/D conversion is intented from 0 % to 110 % of the allowable analog signal, resulting in a number from 0 up to 30000 (and counting), where in fact we use the 0 .. 27648 bandwidth as 0% .. 100%

I can accept that in industrial use this kind of conversion is generally acceptable. I can accept that in the same numerical tag there are also the main diagnostics informations. But I can't accept that the card are selled as 16 bits, where in fact the resulting data is treated as 14 bits.

I much prefer the solutions from other manufacturers that provide diagnostics on bits outside the conversion word, but do not cut the data, perhaps forcing the programmer to write the filters and various offset corrections himself, but without losing anything in the definition of the 16 bits, which in some circumstances remain welcome.

Is screaming heard in the middle of a thunderstorm?

My opinion about this case is the same as yours. I am less bothered and less vocal about it. Luckily our company has the flexibility to use other brands when the requirement demands it.