Continued from:
http://www.mcselec.com/index2.php?option=com_forum&Itemid=59&page=viewtopic&p=63514#63514
I played a long time ago with various S&H circuits, and always found that the charge injection from switching FETs gave far too many transients. Nowadays you can get CMOS switches which are orders of magnitude better than things like the old 4066, and individual FETS. Maybe these would be better. DG300 rings a bell.
Certainly switching a couple of switches into capacitors using a couple of lines from the micro - that part is easy, any time delay you like, from microsec to months. But because the delays are different, any transients will also be different, and so you will still see their effects when you do an ADC conversion on the result.
Regardless, you will still have problems with the mechanical attributes of solenoids. Notice the noise on the real current plot in that first paper. I will bet that much of that is caused by mechanical drag as the plunger moves inside the body. Each solenoid will have that, and it wont be consistent in time, or with load, or even from stroke to stroke. Working out precise positioning in the face of that might be a hard job.
There are extremely good and cheap position sensors based on optical or magnetic techniques which will give precise measurements, without relying on the internal characteristics of solenoids. For instance have a look at how those digital calipers work. I would investigate something like that first.
↧