Home > Hardware > DATAQ Starter Kit Support > Forum > Is discontinuity in data logging anticipated in DI-155 ?

Is discontinuity in data logging anticipated in DI-155 ?

Oct 17, 2014
jagdish wrote
I recently procured a DI155 USB data logger with WinDaq High speed. However when the equipment was tested for long duration data logging (at 10 Hz sampling frequency), it is seen that the data logging takes place correctly for sometime after which the logged value does not match the actual voltage present. In fact in all the trials. the measured value drops down to a smaller value which remains constant. This was tried with two different computers and the results were the same. Voltages measured at the analog inputs comply with expected values. So the problem lies in one of the following
1) The analog-to-digital conversion circuitry in DI155 could have issues operating over long durations due to possible issues like heating etc, which could result in wrong data conversion
2) The USB circuitry could have issues which could result in discontinuous transmission of data to the computer.
3) The issue could lie with the USB cable provided with the starter kit (Alternatte USB cables don't work)
4) There could be come sort of buffer roll-over in the DI155.

Could you please clarify what the possible issue could be and how to evade the same?

Thanks
Jagdish
Reply
5 Answers
Oct 18, 2014
Ken agent wrote
Does WinDaq continue to move the erase bar or scroll the display after the value becomes constant? Does the value at left of "% used" continue to increment? Is a "Gaps:" count displayed to the right of that field?

Although you may be recording at only 10 Hz, if Edit Preferences Maximum Sample Rate is set to its default 10 KHz, WinDaq samples at the higher rate to compute averages of 1000 samples and may fall behind on a slow computer or busy system.

On Windows 7 and later, UAC Virtualization can contribute to such problems. For a device using port COMxx, the need for this can be eliminated by copying files DI1xxNT.DLL and DICOM1xx.EXE from C:\Windows\SysWOW64 (or C:\Windows\System32 on 32-bit systems) to the installation directory where DI100ANT.EXE resides. However, if plugging the device into a different USB socket subsequently changes the COM port number, this fix will fail until files for the new COM port are copied.

Make sure that neither of Edit Preferences "Scroll Limit" or "Write Through" are checked.

If you were sampling at 10 KHz instead of 10 Hz, you should try the suggestions in this paragraph. Avoid using Options Scroll Mode. Increase Edit Compression and Edit Preferences Skip Factor to avoid using all samples for compression, which forces acquisition compression to a multiple of the skip factor, but allows WWB to open the file with the desired compression. Increase Edit Preferences Update Interval (which makes the display jerkier). Shrink the WinDaq vertical window size. If not being able to see the waveforms is acceptable, check Options Pause Graphics in lieu of the other graphics options in this paragraph. This still updates the "% used" value and any gap count to indicate recording progress.

If the problem persists when using an acceptable combination of the above suggestions, unburden the system by eliminating competing activity such as simultaneous Internet access, instant messaging, checking for email, antivirus scans or updates, or network access by other computers.
Oct 21, 2014
jagdish wrote
To answer some of the questions asked
1) The scroll bar still moves after the value becomes constant. The memory occupancy still increases. There are no gap flags during the process

2) This is the case even if the Edit/Preferences/Maximum sampling frequency is set to a lesser value

3) Copying the specified .dll and .exe files to the installation folder does not work either. This was tried out

4) The remaining suggestions point to steps to be taken if the maximum sampling frequency is 10kHz. Since this maximum sampling rate was reduced, these are now irrelevant

Few more things...
The constant value displayed/stored in place of the actual value of voltage actually changes when the voltage range is programmed to a different value, which makes no sense if it the voltage is actually being measured. So I guess there could be a communication issue between the device and the PC.

Please provide possible suggestions

Oct 21, 2014
Shawn agent wrote
Jagdish,

What are you connecting to the DI-155? What is the source?

Regards,

DATAQ instruments Suport
Oct 23, 2014
Ken agent wrote
In the hardware manual di-155-manual.pdf, under Specifications, Analog Inputs note:
Maximum common mode voltage: ±20 V peak

Since the constant value displayed/stored in place of the actual value of voltage actually changes when the voltage range is programmed to a different value, I suspect that the problem arises because as acquisition proceeds, the common mode voltage of at least one of the inputs with respect to USB shield or computer chassis ground drifts outside the ±20 V range. When this occurs, use a high-impedance DVM to measure those voltages. If while making that measurement the displayed value changes significantly, that would also confirm this suspicion.

To solve the problem, connect the ground return from the signal source to an unused negative digital input, which is connected to ground through a 4.7K 0.1W resistor. If the common mode voltage still remains larger than ±20 V, that would exceed the power rating of the internal resistor.
Oct 24, 2014
Ken agent wrote
If a DVM does not show a strong DC common-mode voltage, switch it to measure AC voltage. Sometimes line filter capacitors in a charger powering a laptop with an ungrounded chassis will cause the chassis to float at half the line voltage, but not with enough current capability to be noticed or dangerous. Connecting such a laptop to a signal source referenced to earth ground would be troublesome. A common mode voltage of 14.14 VAC has peaks of ±20 VDC, which would exceed the common mode range even with small input signals. Disconnecting the charger or grounding the laptop chassis would cause the problem to disappear.

If you really need to measure signals in the presence of strong common mode voltages, see www.dataq.com/products/di-245/ for a device that can work with 120 VAC common mode and excellent common mode rejection. However, its sample rate is limited to 2 KHz with one channel enabled, or 200 Hz total with two to four channels enabled.

Since the erase bar keeps moving and no gaps are being reported, you should delete any DI1xxNT.DLL and D1COM1xx.EXE files that you copied to the application directory. Leaving them there could cause problems after installing future software updates, since the files in the application directory would be used instead of the updated ones in the system directories. For optimum noise rejection by signal averaging, you should also raise Edit Preferences Maximum Sample Rate to 10 KHz and use File, Save Default Setup.