OP
OP
Borat
Valuable Member
What is "Order from Pak" and is it a curse or a treat?
DIY Robotic Tester - BoraTronic
- Thread starter Borat
- Start date
- Tagged users None
Ben Pedersen
Valuable Member
- Joined
- Jun 7, 2018
- Messages
- 1,050
- Reaction score
- 953
Minhaj Qazi
Community Member
consider it a treat rom pakistan.
I do not think PWM is a good idea in this context. PWM is OK for the human eye which integrates light over a period of time. But the RGB sensor will pick momentary values, at which time the LED can be either ON or OFF. Therefore you will be measuring erratic values. Taking the average over multiple readings might help a bit, if the interval over which the measurements are accumulated is long enough, but there will be no benefit from using PWM anyway. I think you should use a simple white LED supplied from a stable voltage source through a resistor. Or, if you want to control the current for some reason, use a linear current regulator, not digital. You can find the schematics based on commonly available inexpensive components if you dig the Internet. Or drop me a line if you need help with this.
OP
OP
Borat
Valuable Member
May I ask for your advice here:
what kind of LED are you using to shine light through vial?
On my PCB design I currently have (a number of) power connectors for 3.3v, 5v and 12v - but I am wondering if I need some other (lower) voltage regulator for this led... The LED does not need to be super-bright, I think something like 1W led would do?
OP
OP
Borat
Valuable Member
Ok folkes, following my version 1 blunder (some erroneous choice for Arduino Mega 2560 pinout) , I have re-designed the PCB. This time I printed it on a piece of paper and ensured that:
(1) components' pins layout match the design
(2) components do not obstruct each other
This time I also added various redundancy connectors:
(1) additional 5v voltage regulator with output pins
(2) 2x additional 3.3v voltage regulator with output pins (12v->3.3v and 5v->3.3v)
(3) additional unconnected 3 pin and 4 pin connectors
(4) plenty of "spare" connectors, some of which will be used for controlling optical switches to position the rig
PCB is ready for production, I will upload it tonight for printing in China.
(1) components' pins layout match the design
(2) components do not obstruct each other
This time I also added various redundancy connectors:
(1) additional 5v voltage regulator with output pins
(2) 2x additional 3.3v voltage regulator with output pins (12v->3.3v and 5v->3.3v)
(3) additional unconnected 3 pin and 4 pin connectors
(4) plenty of "spare" connectors, some of which will be used for controlling optical switches to position the rig
PCB is ready for production, I will upload it tonight for printing in China.
bro..
your on another level
your on another level
Garia666
Active Member
Indeed I used to follow the dude on utube but he discarded the animals to quick for my taste.
OP
OP
Borat
Valuable Member
No need to discard animals - they can be a tasty snack!!!
Garia666
Active Member
Lol your like a genius clown
so when will this be mass produced so you can take over the world.1W LED would be way too powerful, you would probably need only 0.1W. But you can use a 1W emitter and drive small current through it - that way it will serve longer and produce much less heat. Forward voltage over white LEDs is about 3V (for good LEDs Vf should be something like 2.6-2.7V, but higher for crappy Chinese LEDs).
So, for 1W LEDs you can drive a current up to 300mA. But if you drive up to 30mA, then it will become 0.1W LED. The 5V supply should be sufficient, but you can also use the 12V supply.
If you want very fine current adjustment and stability, you could use something like LT3092 - but there are no ready-made modules using this chip. Of course, you can design a small PCB for this, but it will probably be an overkill. Providing that your power supplies are stable, I suggest that you use a simple resistor connected in series with the LED. You can calculate its value using Ohm's law: R=(Supply_voltage - LED_voltage)/Desired_current. If using 12V supply, and supposing that your LED Vf is 2.7V, and you want 30mA (0.03A) current, then the required resistance would be (12-2.7)/0.030=310 Ohm. Meaning that you can use a standard 0.5W resistor of either 300 Ohm or 330 Ohm, connected to your 12V supply in series with the LED.
If you want to use the 5V supply, you will need R=(5-2.7)/0.03=76 Ohm, so you can use the nearest standard value of 75 Ohm (next is 82 Ohm and will give you slightly less current). But in case of 5V supply, any voltage instability will result in more variation of current if using this circuit. So, if you go with the current limiting resistor approach, I suggest that you use the 12V supply, and if you want to use the 5V supply, you'd rather use a chip that is capable of precise current regulation (like LT3092 that I have mentioned above)
So, for 1W LEDs you can drive a current up to 300mA. But if you drive up to 30mA, then it will become 0.1W LED. The 5V supply should be sufficient, but you can also use the 12V supply.
If you want very fine current adjustment and stability, you could use something like LT3092 - but there are no ready-made modules using this chip. Of course, you can design a small PCB for this, but it will probably be an overkill. Providing that your power supplies are stable, I suggest that you use a simple resistor connected in series with the LED. You can calculate its value using Ohm's law: R=(Supply_voltage - LED_voltage)/Desired_current. If using 12V supply, and supposing that your LED Vf is 2.7V, and you want 30mA (0.03A) current, then the required resistance would be (12-2.7)/0.030=310 Ohm. Meaning that you can use a standard 0.5W resistor of either 300 Ohm or 330 Ohm, connected to your 12V supply in series with the LED.
If you want to use the 5V supply, you will need R=(5-2.7)/0.03=76 Ohm, so you can use the nearest standard value of 75 Ohm (next is 82 Ohm and will give you slightly less current). But in case of 5V supply, any voltage instability will result in more variation of current if using this circuit. So, if you go with the current limiting resistor approach, I suggest that you use the 12V supply, and if you want to use the 5V supply, you'd rather use a chip that is capable of precise current regulation (like LT3092 that I have mentioned above)
Last edited:
OP
OP
Borat
Valuable Member
I am inclined to just use the led onboard the chip: I can make reflective surface on the opposite end easily and see if that' sufficient...
Yes, you can do that, and it will probably work. But the detector will also catch the reflected light from the vial wall - and that light has not passed through the solution. Yes, you can use cancellation for this, as it will probably be a constant, but it will reduce the dynamic range. In my opinion, moving the LED to the opposite side to the detector is a better approach. But of course, you may start it as you mentioned, and then can fetch another LED on the opposite side at a later point.
Similar threads
