06-23-2014 07:02 AM
We are looking for the UDV (Ultrasonic Doppler Velocimetry and not a User Defined Variable!) solution based mainly on NI hardware and a PC with running LabView. Except for the UDV transducers, which can be purchased elsewhere, I wonder if somebody can recommend a DAQ card or special "re-programmable" processing unit for acoustic signals which have been successfully employed for UDV measurements? By re-programmability I mean to have a possibility to tune some parameters such as timing and shape of the emitting pulses, emitting frequency, pulse shape and repetition cycles, few channels etc. Another desirable (though optional) feature would be to have a frequency sweep function allowing a spectroscopic probing of the media for its resonant acoustic response. Beside high cost, the re-programmability is actually another reason why we are considering NI based solution as an alternative to ready-to-use UDV unit.
Maybe somebody has used FPGA for UDV measurements? Any recommendations for the hardware used?
Skipping boring details about our scientific research goals and motivation, the primary task is to measure velocity (ranging 0.05..0.5 m/s) and distance (ranging 0.04..2 m) to the gas bubbles floating in a fluid having high density (with sound velocity of about 1800 m/s). Note, the real-time signal processing is not required! The signal emission, reception and recording with following post-processing on PC is typical approach for us.
All suggestions are appreciated!
07-11-2014 03:31 AM
When choosing DAQ card you need to know what type of signal your sensor deliver to be sure how to measure it and with what (Analog voltage/current, digital, pulses (counter)).
When you now what type of signal, then start looking at what type of board you need. Things to consider then:
-Number of channels/counters/DIO.
-form factor (USB/Ethernet/PCI/PXI)
Our if you need a complete custom acquisition with a FPGA board.
Test the solution help:
07-11-2014 08:03 AM
Some additional thoughts....
Frequency range? US could be 40kHz or >1Mhz .... migth depend on the damping of your fluid 😉
0.04(m)/1800(m/s) means 2µs response so I assume something more in the MHz
RX&TX with one device? Fast mux or blanking needed? (Never saw a overload recovery time specified in the ni cards .... )
Sending power .. impedance matching ... looks more like a AWG with 50 Ohm output , direkt feed or is an amplifier included in the transducer? cable losses?
Expected output of the receiver? conditioning needed?
07-14-2014 03:06 AM
Thanks for your tips and right questions you have asked!
I agree that upper bound in frequency will be 8-10MHz, 50 Ohm matching impedance, we need to send/receive voltage pulses. Sure, lots of parameters are depending on transducers we will use, however currently I am not certain what transducers we will buy. This is because our fluid is a liquid metal (200-250C working temperature) and unfortunately is very corrosive (thus transducers lifetime is limited, but that's ok).
I think PCI card with at least 2 pairs of channels (RX+TX) would suit us. Though, a stand alone unit with Ethernet connection will also work. USB/cDAQ might not be sufficient in terms bandwidth (if fast multiplexing of the channels is applied), and I would not like to mess with PXI for now but there is a white paper solution, have anybody tried something similar?
Would be great if somebody can give a working example regardless of the fluid type/application used and, most importantly, share some pros and cons of that solution. A complete non-NI based UDV solution would cost somewhere in between k$20..k$40. With NI hardware my price estimate is about 50% excluding some programming man-hours. These 50% saving for our academic institution might sounds good, but here in Sweden man-hour is not cheap 😉
The cost is not a major issue here. In our application we would like to have some freedom in adjusting some parameters. For example, most of the stand alone UDV units are preset to a fixed sampling frequency (for example vendor can set either 1; 3 or 10MHz). Here is an advantage to have NI DAQ card. Another feature which we would like to have is to output customly modulated signals instead of preset pulse. Some LabView programming will certainly extend our measurement techniques.