A bit ago I got the idea of trying to determine the mixture of coolant vs. water by detecting both light occlusion of the mixture and the wavelengths blocked / passed. This initial device is just a rough prototype that will assist me in determining a course of action in regards to overall design.
The idea behind the device is to have the sensor fitted on to any pipe attached to the machine where coolant goes through and give a live measurement of coolant mix and alert the operator if the mix gets too high or low. Also I’d like it to detect tramp oils that have been beaten in to the coolant.
The board itself is pretty simple, just a PIC16f690 hooked up to an rs232 driver and using three analog channels. I may in the future build a more sophisticated ADC board, but for now, this will do. The mcu is linked to the two rail to rail opamps, 7014D’s to be exact. they were needed to condition the signal from the LEDs.
The sensor area is basically a cup with a white LED as a light source for the sensor LEDs. The three LEDs are IR, Orange-red, and Green. The LED’s respond to wavelengths more energetic than the ones they emit, therefore, the selections I made. I did try a blue LED but the response wasn’t good at all. Hopefully it will provide useful data, hopefully.
In order to make data easier to collect, I put rs232 communications on it. I can store and track data this way. Above is some of the terminal output. Notice that I’m taking 10000 samples… this has the effect of increasing, to a very limited degree, the resolution of the device. It is however fraught with error thus far…
Anyways, any data collected and the design of this is extremely preliminary. I’m not even sure it’s a valid or useful idea yet.
As an extra bonus, or punishment, here’s a video I made for this device.
Well, I’ve been designing some new stuff since having a little wind of inspiration. This head weighs in at about 700g and provisions have been made to allow an attachment to a handle and therefore a hand-held version could be made. Here are some images.
I’m going to use these design concepts on the next in-process version. I’m going to design the external case for the electrical portion and for possible battery storage. Since the motor only draws 100ma at 24v or 200ma with a 12v motor, a hand held version is definitely possible. This version will use a PIC16F767 for control and using a smaller LCD display to display info.
Now, on to developing some other devices. I’ve wasted far too much time in the last many months, time to get busy.
Well, I’ve been looking around at torque sensors and found some interesting information. For one though, they are NOT cheap. Keep in mind that these are higher sensitivity sensors but it doesn’t seem to have an overwhelming bearing on the overall price. for example.
Anyways, I’ll keep looking for a cheaper unit, though, I doubt I’ll find one cheaper than the unit I developed. Perhaps I’ll machine a small enclosure with precision bearings and develop a more robust and practical design. I have some conductive graphite for the brushes in order to reduce noise and the brushes could be replaceable. Might try it.
It hasn’t been long since my last update about this thing but I’ve made some progress as far as making the unit into a single board. I simply used the AD7705B as I suspected it work work great for this application. I’ve been playing around with reducing the amount of noise coming from the sensor but random fields are causing the ad7705 to produce strange numbers, even when my hand just gets close to the unit. Here are some pictures of the unit.
Thus far it seems to work though, at least when tested with a multimeter. Sadly though, it seems impossible to test it with the oscilloscope because it uses the same ground as the power supply, damnit!
Well, so far, so good. That’s all I can say at the moment. The rotary gage is operational, though it hasn’t had brushes made for it and it seems to output a reasonable amount of voltage difference based on displacement.
Here’s what it looks like so far:
I’ve put it through its paces thus far and I can see a 10mV difference when I move the sides, that’s good enough. Soon I’ll see if it works with the stuff attached.