Pictures
Pictures
Here are some pictures of various devices I've built back in summer. I think there were eight in total!
Photobucket Album
Many folks were asking how the first prototype was doing. Well, you can see it right there - it's the big white blob on the left. I still have is somewhere in my garage...
The most advanced one was the "Grey Box". It was based on SIA principles and worked great. The only disadvantage was the cost of the two major components - syrynge pump (Cavro XP 3000, $1,400) and the selector valve (10-port Rheodyne LabPro EV750, $1,300). Luckily I found all the parts on ebay. Cavro's are easy to find but I sincerely wish you a good luck finding another selector valve like that. There were total three stepper motors in that grey box. It's definitely not a toy or a school science project...
The photodetector was homemade and extremely cheap and the rest of the electronics was less than $150 I guess. This is a pretty nice and flexible chem analysis platform for a well funded lab but not a hobbyist. Several companies out there are making these devices and one can buy it for ~$15K (it will also need a PC, software and a decent spectrometer - another $3,000-$5,000).
Here's a good example of a versatile and compact SIA device. Keep in mind while looking at it taht my device had everything inside the box:
http://www.oceanoptics.com/products/fiasialov.asp
Then I attempted to simplify the design by ditching the uber expensive selector valve for solenoids and limiting the device to a single reagent and a single source (three solenoid valves) and swapping the precision syrynge pump for an FMI pump (stepper with a white head thingy on most of the photos). It worked quite well at times and it took me almost two months and four prototypes to get the FMI pump into the right position so it will not choke on air bubbles and what not. The resolution of the device was limited to about 2% by the pumps I had in my posession - they were delivering 5uL strokes at best. The pumps with better resolution are available from FMI in OEM shape for "only $600 a piece" but you can always get one on eBay for $10 in a decent shape. In the end I got sick of FMI pumps drifting and buying pumps off of ebay was not an option for a production run anyway.
This got me thinking again and the result was a homemade syringe pump which costs less than $100 in off the shelf parts and is rather simple to produce. Stepper motor provides the necessary precision and stability and even worst plastic syringe I had at home survived 10,000 strokes without any leaks - we need the pump to perform about a thousand strokes per year if we wish to monitor Alk three times a day. For reference, Cavro pumps can do 1,000,000 strokes before you replace the syringe ($150 each). We don't need that level of durability at that price (>$1K here).
Also at that point I figured that there is a way of determining the endpoint of the titration by monitoring conductivity. It's much more tricky than watching colors changing from blue to red, but it worked quite well in first approximation.
Conductometric titration method was better because there is no need for the pricey reagents (bromocresol green/methyl red) and the whole detector can be made on a printed cirquit board, including the gold plated electrodes. Much better than messing with some blue-red LEDs and photodetectors and drilling 1mm holes in plastic at various angles. The four-electrode design was chosen as the most advanced and it took me another three months (until December) to figure out the best way to implement it and learn how to design PCBs with SMT components and stuff.
A week ago I was ready to place an order for the PCB fabbing but for some reason was hesitating to spend $500 at once and decided to give a google another chance and searched for some alternative methods of measuring conductivity again. Turned out that there is indeed a much more advanced way of measuring conductivity of solutions discovered back in 80's that becomes more and more popular among analythical scientists these days. They use capacitively coupled contactless electrodes and it involves high frequencies and high voltage excitation. Apparently this should work for what we need it here. They can measure a very slight variations of conductivity in a microbore tubing (30um ID) through the wall (365um OD) without touching the liquid at all. This must be the best method as the electrodes will not get corroded or fouled over time and our goal here is to build a device that will last for years without service.
...just got the necessary wideband Burr-Brown opamps in the mail today and will try to breadboard this method over the weekend. Wish me luck! 24V will be the highest I will go anyway..
-Kyryl
