The data acquisition system used in this series of experiments was composed of all three main systems used at the OSU GTL because of the large number of channels.
They will be described briefly below as individual systems, and then the overall integration will be described as well as the main calibrations performed.
Main DAS – This is the 256-channel high-speed data system produced by DSP Incorporated. It is a CAMAC based system controlled using computers via GPIB
(General Purpose Instrument Bus, or as others know it, HP-IB). This system is composed of 256 channels of separate amplifiers and digitizers (all fully programmable). The overall system runs at rates up to 100Khz. The digitizers are 12 bits ±5 VDC range, and the amplifiers have full differential input and provide offset injection, programmable gains and 8 pole anti-alias filters. All channels are simultaneous sample and hold and when the entire system is running, 256Ksamples per channel are possible. As fewer channels are turned on, the memory for the active channels increases. This system also houses the spincoders that record the rotor position using the encoders, and measures the time of each encoder pulse (500 or 1000 per rev) at a 10 MHz rate. This technology is relatively old (from the 1980’s and 1990’s), but still is one of the quietest systems around, generally out producing in overall system specifications many newer 16-bit systems at equivalent gains.
27
High-speed System- This is a 19-channel Datalab system, made by Lucas Instruments in England. This system dates back to the late 1980’s and is similar in construction to the DSP system (with dedicated amplifiers and digitizers) but it can acquire data at rates of 1MHz, although it works better at 250 – 500 KHz. The amplifiers, while programmable are not of the same quality as later instruments so these are most often run at relatively low gain, and a dedicated pre-amp is used (currently a Vishay 2210 system). This system is extremely versatile, but under general use, 64Ksamples per channel are possible.
High Channel Count/Low Speed System – Currently two National Instruments 64 channel E-series boards are used to provide 64 channels (total) of differential input signals. The maximum A/D rate for each board is 1.25 MHz, although these are
multiplexed channels and not simultaneous sample and holds, so there is some skew time between channels. Generally, these are run at 5Khz acquisition rates with channel rates of 200 KHz and are used for thermocouples and RTDs where generally only the time- average properties are of interest.
Together, these form an A/D system with 339 available channels. The entire system is controlled either via X-11 code or LabView code. The original X-11 code is slowly being phased out in favor of the LabView code, where the main analysis is done.
The data is stored in a binary format. Typical run sizes are on the order of about 35 Mbytes, and each channel generally has 20, 000 samples (or about 0.2 sec of data at 100 KHz). The system also includes several other pieces of reference equipment such as Keithley 2001 (7 1/2 digit) multimeters, HP 322130A Wave generators, and EDC 522 Precision voltage sources, used in the calibration systems, and trigger systems, all of which are computer programmable and set-up in an automated test environment.
All instruments requiring power (aside from thermocouples) have individual signal conditioning, all built in-house at the OSU GTL. These signal conditioners consist of power-supplies and pre-amplifiers depending upon the sensor type. For the Kulites, these signal conditioners have unity gains, but extremely high bandwidth (1% drop-off points at about 400 KHz). The heat-flux signal conditioners all have constant current supplies that deliver 1 ma at frequencies also in the 100’s of KHz at a 1% reduction in amplitude. In addition, these also have gain capabilities. A picture of the signal
28
conditioners (preamplifiers and power supplies) is shown in Figure 3.6. The left side contains the power supplies for all the Kulites. The right side has the amplifiers/power supplies for the heat flux gauges. The center section contains the main patching to the A/D systems. This picture was taken immediately before entry 2.
Figure 3.6 DAS System Preamplifiers/Power Supplies and Patch Panel
Figure 3.7 Main DAS systems
Main DSP System Datalab System National Instruments
29
A picture of the DAS systems is shown in Figure 3.7. One can easily see the DSP system and the Datalabs, the National instruments system is behind the far corner,
although one can see the computer system that houses the acquisition boards. Generally, two to three computers are used to run this system. A third computer runs the Siglabs that provide real-time FFTs of the accelerometers, used to monitor the condition of the rotating system bearing package.
As a final topic, it is worth discussing the primary calibrations used for these experiments. All pressure calibrations are done relative to a Heise HPO 150 psia pressure sensor that has a rated accuracy of 0.05%. Two of these sensors are kept
operating at all times to provide a check on each other. These Heise sensors are routinely sent out for calibration and maintain their rated accuracy for well over a year. The heat- flux sensors were calibrated to obtain the resistance versus temperature characteristics using a laboratory grade Azonix platinum RTD probe in an oil bath or a vacuum oven.
Pressure calibrations start with each airfoil being calibrated after the installation to check for hysteresis and non-linearity. Once the rig is assembled, the main pressure calibrations are done as often as needed (usually once a day) with the entire system in place by slowly pressurizing the entire test section to about 60 psia (or whatever the vane inlet pressure will be for the experiments being performed) and then slowly venting. This provides an excellent verification that all channels are operating over the course of the experimental program.
The heat-flux gauges are calibrated more strenuously before the experiments since there is currently not a good way to perform an in-place calibration similar to the pressure sensors. Each gauge is calibrated before it is installed in the oil bath, and then the airfoil is calibrated once again in a vacuum oven after the installation to verify that the gauges have not changed operating conditions. Then there is a check on the overall system to get the proper lead resistance. The resistances for each sensor are measured and the
preamplifier settings checked before each run so that gauges that undergo large erosion can be monitored. At the end of the experimental program, gauges that could easily be recalibrated are recalibrated. Currently, we are working on a method to recalibrate all heat-flux sensors in a post-measurement program.