So this year, we’ve decided to use Banner’s QS18VN6D optical sensor to detect the presence of balls in our acquirer. Sadly, it doesn’t work as of yet.
We know that the sensor itself works for two reasons
1 - the yellow light turns on and off in response to an object in front of the sensor
2 - when the oscilloscope is hooked up as per the included diagram*, it reads high and low values (the white lead is low by default; it turns high when something is in front of it)
However, the problem is that the signal bizarrely doesn’t exist
1 - when a multimeter is hooked up to read the circuit as indicated*, the voltage across signal and ground is a constant 0
2 - the computer sees constant “TRUE”, likely because of this^ problem.
When on the robot, the power supply is a 12V, 20A output from the PDB; the signal is hooked up to, obviously, the Digital Sidecar, by connecting the “sensor in” as seen here* to a Digital I/O.
http://dl.dropbox.com/u/11542623/Banner.png
*Sensor in would be the signal lead on the scope/pos lead on the meter, sensor out would be the ground leads on the scope and meter.
Does anyone have any ideas as to why this isn’t working**? Thanks.
** On an afterthought, I just remembered that at one point, we accidentally hooked up the Digital I/O pin from the sensor to the 5V power instead of the signal (it is now appropriately plugged into the signal pins) - could this have fried the sensor, somehow? I doubt it, since the sensor operates up to 30V, but I’m just tossing it out there.
Based on the datasheet you linked to, this is an NPN (aka sinking) output. You would want to connect the collector to your digitial I/O line so the sensor pulls it down or lets the pullup resistor pull it up.
Yes, there may be some minimum voltage it operates at, and 5V might not work. However, based on your schematic, I think you have it hooked up wrong. Refer to the wiring diagrams on page 5 of the datasheet. You want to apply 12V across blue and brown. Then, using a suitable pullup resistor (the digital sidecar has pullups on its inputs), connect to either black or white. You should see the state of that signal change when you present an object versus when an object is not present.
You can use 5V on the pullup (the open collector output will just pull it down), but you need 12V across blue/brown to give the light and detector circuitry the appropriate supply voltage.
I wired in a pullup resistor as per the spec sheet diagram across the white lead (as indicated in the diagram I posted). The power supply is 12V, as the operating voltage of the sensor is 10-30V (as I also noted).
Thank you for your input, although those problems you point out are not the ones I am having.
Jake, that issue has been resolved, but I can’t be sure if such a hookup would or wouldn’t have lasting effects on the sensor itself.
If you can describe a bit more of that scope flicker, it may lend an explanation.
Here’s a possibility: If you get rapid flicker when the object is present, then it’s probably a supply issue where the system can operate with little load (i.e. it’s not pulling current through your 1K pullup). Once it has to switch, however, the current draw causes a power dropout. This could be either the internal circuitry going undervoltage or your supply line going undervoltage. Once the unit loses power, the supply returns to the nominal condition, it turns on, and it tries to signal again, repeating the power loss. This would explain a fast flicker with an object constantly presented to it.
As your teammate said, verify your supply voltages and report back.
As a cautionary note, I’m not sure if the DSC is 12V tolerant. Those 5V pullups on the DSC PCB would do the job you want.
I hadn’t realized the implications of that last comment in your first post.
Referring back to my earlier comments about hooking it up wrong, I was referring to your application of 5V to the sensor line.
When that connection was made, you hooked up voltage directly to the collector of the NPN transistor at the output. When the sensor switched to indicate the object was present, there was no resistor to limit current.
So, yes, you may have fried the output by exceeding the power dissipation spec of the output transistor (which isn’t in the datasheet…but I’m sure is not large). If you have an untouched sensor, try hooking it up and see if you get the same result. The flicker you observe could be that potentially damaged transistor.
Sorry if I wasn’t clear on that point; by “flickering” I meant to say that the sensor appropriately responds to the presence of something within its field of vision by turning on and off. The latter issue has since been rectified (it was inducing an output overload signal, which is how I noticed it).
The RightSight sensors last year were wired in a similar fashion; their power was supplied by the PDB, and their outputs directly into Digital I/O, so I’m inclined to believe that this isn’t necessarily the case, though I could be wrong. (In that event, would I have to wire in voltage converters or something similar? - that seems excessively complex to me).
What exactly do you mean by the 5V pullups on the DSC PCB? emphasis mine.
Don’t use a resistor when connecting it to the robot. Just give it 12 volt power on the blue wire, common on the brown one, and run either white or black to a digital input signal pin on the Digital Sidecar. There’s already a pullup to 5 volts on that pin, and it will read the sensor’s output just fine as it is.
PCB = printed circuit board. By that I refer to the guts of the DSC, if you removed the blue housing. As Alan pointed out, there are already pullup resistors there on that board, so you shouldn’t need anything external.
So you hook the Brown+ and Blue - up to the 12 V powerdist. Board with a 20 Amp breaker and then you connect the white wire to the white wire of a PWM and then to the digital Sicard
Blue -> ( - )
Brown -> PWR
White or Black -> SIG