HF Scientific Turbidimeter Support
Importance of Getting the Detailed Data
The MicroTOL Turbidimeter manufactured by HF Scientific is used to monitor the influent and effluent turbidities of the laboratory flocculator setup. In the current experimental setup, we are sampling turbidity measurements from each turbidimeter in 1 second intervals using Process Controller.
The MicroTOL Turbidimeter is a nepthalometer that measures turbidity by sending a beam of light through a sample volume and measuring the intensity of light across the sample volume at 180 degrees and at 90 degrees to compute the amount of scattered light. The light intensity readings measured by the optical sensors are then transduced into an analog voltage signal that can be understood by a microprocessor that can then compute and display a NTU reading. We suspect that the sample frequency of the analog voltage data is much faster than 1 Hz, and that the reported NTU values are an averaged value integrated over one second. We suspect that this onboard integration has averaged out some important information that may give us a more accurate picture of the size and velocity of the flocs passing through the sample volume.
By looking at the analog voltage data, we can decide how to filter the data in a way that will minimize background noise but give us good temporal resolution of the change in turbidity in the sample volume. We should be able to draw better conclusions about particle size and sedimentation velocities this way than from the analyses using averaged data. We are currently in contact with HF Scientific to determine how to obtain this analog data.
Contact Log with HF Scientific, Inc. Technical Support
Friday, February 15th
Gemma talked with Phil/Joel in technical support. It is possible to get the "Raw Data" and we can do this by incorporating MODBUS communication programs. Phil does not know the exact timing of the analog signal, but he believes that is as low as 250 milliseconds (maybe even lower), however, he also said that we might not be able to get a clear enough (readable) signal until we reach 750 milliseconds. He will figure out the exact number and respond next week.
To read the raw data, we need to user Hyper Terminal, a Windows program that can bee accessed under Windows>Accessories>Communication. The HyperTerminal manual should be read to figure out exact settings/configurations to read from Modbus. Next, we have to buy the Modbus program package (or this might come free with HF Scientific software).
Monroe's comments: This is all digital data. We want analog data. We want voltage, not serial communication using MODBUS. We are already using the modbus (not sure on this terminology, but we are definitely using RS485 communication protocol) to get the data from the instrument. And I think we are already getting a data point every second. The process controller software uses serial communication (in a more sophisticated way than the hyperterminal approach) to get data from the turbidimeter. If we can get an analog signal then the meaningful data rate will only be limited by the detector (sensor) response characteristics. Check out: Floc Strength Measurement
Wednesday, February 21
Spoke with Joel and Mike Goodman (Director of Technical Support). There are two ways to get the raw voltage data.
1. Obtain the voltage before it goes into the ADD. These numbers will be 0-3 Volts and we will need to associate the voltage with an NTU number. We could also try hooking up a data logging voltmeter to this.
2. Download a version of MODBUS POL, a shareware program that will get the voltage data converted into NTU data. The program can get data every 150 millisecond, but using MODBUS POL, we will only be able to retrieve as fast as the HF can update it, which is approximately 250 millisecond. The digital numbers can be dropped into csv file and into excel easily. Joel was saying that we might not need MODBUS POL if we have a program that is retriving data (Process Controller). I think we need to distinguish the functions of MODBUS and Process Controller to see if PC can accomplish what we need.
Mike will be sending us an email with all of these specifics written out on different ways we can retrieve a faster signal.
Monday, March 10
Mike sent the information on how to access the data:
The attached photo shows the connections that will give you the voltage as it is presented to the A/D. These are real time readings that have not been linearized or massaged at all by the CPU. There are logging voltmeters that can be connected to serial ports. I'm sure Cornell has these around somewhere.
The other method that we discussed was to log the data via modbus. We use a program called Modbus Poll. I have provided the link below. The cost is $99 for a single license, but it is free fro a limited period, I'm not sure how long that period is.
http://www.modbustools.com/index.asp
If anyone there knows programming and is familiar with modbus they can write something to get the data.
The modbus addresses (registers) of interest are:
30001 & 30002 Float Type = Sensor reading
30003 & 30004 Float Type = Sensor Raw reading
These are duplicated at:
40031 & 40032 for Sensor reading
40033 & 40034 for Sensor Raw Reading
I have attached a modbus manual for the Micro TOL.
If you need RS485 converters we use B & B (bb-elec.com). For RS485 to RS232 we recommend Part #485D9TB
$74.95 for RS485 to USB we recommend Part # USOTL4 $139.95.
New Raw Data Aquisition
Cameron has outfitted the effluent turbidimeter with the ability to output raw voltage data. We need to figure out how to obtain the data on our PC and understand how to calibrate and handle the data.
LED Lamp
After it was determined that the heat produced from the IR lamp in the turbidimeter was causing recirculation currents in the settling column, the team researched into alternatives. HF Scientific sells turbidimeters with a LED lamp, a no heat producing lamp. After contacting with HF Scientific, the team found that we can simply replace the upper module of the turbidimeter with a new LED lamp. Mike from technical support reported that the LED lamp turbidimeters actually have heat added into the meter to help performance, but that he could disemble this "added heat" function before shipping the module to us.
A recent comparison between IR and tungsten lamp turbidimeters concluded "Turbidimeters with spectral peak responses of the detection systems between 400 and 600 nm give most often lower readings than instruments that complies with ISO 7027 Section 3. Samples coloured by dissolved humic acid give lower readings than corresponding, less coloured ones, especially with instruments that operate in the wavelength range of visible light. For natural waters and suspensions other than formazine, the variation in light scattering properties of suspended particles had various effects in turbidimeters that do not comply with the international standard and the reductions in measured turbidity
were often much larger than those caused by colour. The results of our testing were confirmed in an interlaboratory comparison where two samples were measured by 81 laboratories with various instrumentation."
The new module will ship in May. After installation during the summer 2008, the new setup to prevent premature settling and to prevent recirculation currents in the settling column will be ready and new experimentation can occur.
Back to Tube Floc Home Page