Site Search:
Previous Note Previous Note

DI-145 Data Acquisition Starter Kit Developer's Diary

Next Note Next Note

1/20/2011 Note #11 Ahead of schedule

Sometimes things just work out. We wrote in an earlier post that the revision C board, our first candidate for the first production run, was due in-house by January 24. Things must be slow at the PCB fabricator because the board arrived a full week ahead of schedule on January 17. Three prototypes were assembled that day, and testing has been on going all week. Application engineering and tech support have also had their first look and the instrument has so far performed flawlessly. Some note-worthy items:

  1. The DI-145's analog differential inputs are a tremendous improvement over the single-ended design of the DI-194. A DI-194 analog channel that was enabled with open inputs would tend to drift upward to about a 2.5-volt level, triggering this ubiquitous call to tech support: "I've enabled a channel and with nothing connected it reads 2.5 volts instead of zero. You shipped me a defective unit." Well, probably not.

    Zero volts can only be established by connecting, well, zero volts to the input. The easiest way to do that is to short the input to ground. Sure enough, when customers did that the wayward channel snapped to zero volts.

    With the new differential design of the DI-145, while input signal levels remain within the common mode range of the amplifier (and open inputs surely will), the amplifier's output will be zero or very close to it (along with customer and tech support stress levels). Problem solved.

  2. Ah, crosstalk. If open inputs on the DI-194 settling to 2.5 volts didn't trigger a tech support ticket, crosstalk surely would:

         Customer: When I connect a signal to channel one I see an exact duplicate of it on channel two.

         DI TS: What's channel two connected to?

         Customer: Nothing.

         DI TS: That's normal.

         Customer: What?!

    And more stress. Unlike the DI-145, the DI-194 lacked an amplifier per channel. That meant that the parasitic capacitance of the channel multiplexer could retain a charge from any handy source (like a real input signal), since an open input on an adjacent channel presents infinite impedance. The consequence was a near perfect replica of a connected signal on any open and enabled channels. Disabling or connecting unused channels to ground solved the problem, but not before confusion had run its course.

    An amplifier per channel in the DI-145 design ensures that the multiplexer always sees low impedance, allowing any residual charge from adjacent channels to be discharged before measurement, and thereby reducing if not eliminating crosstalk. Another problem solved.

  3. It seems to us that one of the more popular uses for the DI-194 starter kit was as a fuse. We'd warn customers about the dangers of a common mode voltage. We even included a bright, fluorescent red flyer with every shipment that described how to detect it and how it might, and probably will destroy the DI-194's input channels. Alas, common mode voltages are nearly a fact of life in many industrial measurement applications and our customers and we were fighting a losing battle. No more. Each DI-145 analog channel is designed to survive common and normal mode voltages up ±250 volts, but because of input connector limitations we lowered the spec to ±150 volts. Testing this feature was easy. We simply plugged the analog input channels into the nearest wall outlet. They weren't even stressed, taking the full brunt of ±169 volts with ease. Still another problem solved.

The above three recurring problems, all resolved by the new DI-145, can be visualized in one picture. Figure 1 captures a DI-145/WinDaq real time recording screen. During this session we plugged channel 1 (top) directly into a wall outlet, but left channels 2-4 enabled and open. Then we screen-scaled their displays to show LSB transitions (±20 mV per count.) As you can see, only ±2 counts of noise are on the three open channels, and each is centered at zero. We'd contrast this DI-145 performance with the DI-194, but the latter would be a smoking mass before we could acquire the first data point.

DI-145 Data Acquisition Starter Kit Screen Shot
(click to enlarge)
Figure 1 – Channel 1 (top) has a ±169 V signal applied to it while the three remaining channels are enabled but left open, proving that D-145 analog input channels are immune to crosstalk and large common and normal mode voltages well outside its measurements range. They even stay centered on zero.

The bottom line for today's report is that our first production run candidate has passed all the tests we've thrown at it and the project remains on schedule or slightly ahead. Now we can clean up, verify, and finalize the bill-of-materials in preparation for the first production run. After that, end user software odds and ends as well as internal production calibration fixtures and software need to be wrapped up -- more on that in our next installment.

DI-145 Product Page: