Newbe design question/ Signature analyzer

[uriahsky]

I have been wanting to design a particular piece of test equipment for a long long time.  They do make commercial units but the price seems to me to be way out of proportion to its design. It is very similar to a Huntron signature analyzer but with more features.  ABI Electronics makes what I would like but it costs around $14,000 and I think I can design one.  I have done a few hardware units that do the basics but I would like to be able to do this on the computer. My question is can labview do the following and if so what type of hardware should I use to interface to the computer.  Here is what I would like.  A signature analyze like a huntron that can store the patterns on the computer so I can recall them and compare them to known good boards and chips.  Also the ability to test and store chips both analog and digital in circuit and out of circuit.  I would like to be able to analyze up to 192 channels though the use of clips.  Also to check the current state of a pin on a chip.  If it is high, low, tied to ground, or VCC. and more features as time goes on.  Can labView do this or should I use some other language?   Also if LabView can do this what kind of hardware should I use to tie it into the computer so LabView can work with the data. I need to generate a sine way of different frequencies, adjustable the voltage from 2-15 volts, vary the impedance and use the XY display on a scope.  If I can control those features from the scope then have it generate and save the displays and be able to view up to 12 of them on the screen at any one time.  I have no idea where to begin. Can someone get me started in the right directions. 

Thank you very much

[Kevin_Price]

Honestly, if you have "no idea where to begin", you really aren't ready to tackle all this at once.  You're describing a pretty sophisiticated bit of virtual test equipment and to implement it cleanly and well, you'd need a pretty clear picture of where you're headed before you start out. 

I don't mean this as a put-down.  I'd consider myself a pretty fair LV programmer, especially with DAQ boards.  And for the kind of capabilities you're looking for, I'd consider a $14000 piece of equipment a bargain compared with developing it myself.  It'd probably be iffy even for a quantity of 2.  Remember that the test instrument maker doesn't recoup their cost by selling 1 unit -- they count on selling many units to spread out the development costs. 

Now, that being said, perhaps you can add real value with some extra capabilities.  You'll surely develop your own LV skills considerably in the process which can pay future dividends.  So I'm not saying you shouldn't work toward it at all, just that you might narrow the focus at first, and do some proof-of-concept work on individual portions.  For example:

1. 192 signals -- learn to synchronize hw-timed buffered data acq across multiple boards.

2. Signature comparison -- develop a good file format for signature(s), learn how to compare two signature files to each other, develop a results format that lets you identify the type of location of whatever signature deviations you want to flag.  Will you be looking at specific timing relationships between individual bits?  Looking at at a digital pattern across multiple bits at once?  Cross-correlating analog signals to some window or threshold?  There's a nearly endless # of ways to characterize such a large set of signals...

3. Database of chips - I *think* you'll need to maintain one from the comment about "test and store chips, both analog and digital, both in circuit and out of circuit."  Again, you either need to make a "formal" database or work out a file format to store id data on these chips.  Perhaps the test results must be linked to chip serial #'s too?

4. Generate variable freq & amplitude sine waves 2-15V, vary the impedance.  Don't know where and how you plan to vary impedance, but this isn't a capability normally available on DAQ boards.  Also, most DAQ boards operate in the +/-10 V range.

5. View up to 12 XY charts on screen at once.  Tinker with methods for generating multiple independent instances of a display based on one piece of generic underlying code.  I've used templates for this in the past, but nowadays subpanels would probably be better.  Not sure because that's a feature I've yet to familiarize myself with.

Good luck!  The beauty of ambitious projects is that you'll learn a lot of stuff along the way...

-Kevin P.