I need a voltage sink reality check
Electronegativity
Posts: 311
Suppose I have a Vdd of 6V coming onto a board.
I run that through a regulator that produces Vcc of 3.3V and place an 11 Ohm resistor between Vcc and a Propeller Chip.
Now suppose I run Vdd to the positive side of a blue LED with a forward voltage drop of 4V, and connect the negative side of the LED to one of the Propeller's pins.
It seems to me that the Propeller is sinking 2V, and therefore I should be able to get away with this, but thought it would be wise to ping the forum.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
I wonder if this wire is hot...
I run that through a regulator that produces Vcc of 3.3V and place an 11 Ohm resistor between Vcc and a Propeller Chip.
Now suppose I run Vdd to the positive side of a blue LED with a forward voltage drop of 4V, and connect the negative side of the LED to one of the Propeller's pins.
It seems to me that the Propeller is sinking 2V, and therefore I should be able to get away with this, but thought it would be wise to ping the forum.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
I wonder if this wire is hot...
Comments
Why are you putting an 11 Ohm resistor between Vcc and the Propeller chip?
For the last few years I have been making light up jewelery and giving them as Christmas presents using SX chips and was horrified by their end of life.
I ordered a batch of Propeller chips and need to get up to speed quickly on the new hardware.
There is no problem running an SX with 6V from 2 coin cells and powering LEDs to blink in various pattens, but the propeller will fry if I try that.
The idea of the 11 Ohm resistor was to limit the current to the Propeller to 300mA, but I will leave it out if you don't think it is necessary.
What I plan is a circle of 16 LEDs controlled by a Propeller to display various patterns of light that is powered by 2 CR2032 coin cells.
I ordered some QFN Propellers and found voltage regulators and EEPROM in 6-MLP and 8-MSOP packages to keep everything as small as possible since the whole necklace will only be about 1.8" in diameter. I am using 1206 SMD resistors and LEDs and was hoping to get away without using transistors.
Worst case scenario I have some 2222As in SOT-23 but it will significantly up the part count and board complexity.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
I wonder if this wire is hot...
You should have no problem with red, green, or yellow LEDs running off 3V or less. I've seen some datasheets for white LEDs where the forward voltage is less than 3V at currents under about 10mA and there's a Panasonic blue LED whose datasheet shows a forward voltage under 3V at 2.5mA or less. The forward voltage is very much current dependent and you're not going to want to run these at high currents anyway because of battery life. You certainly don't need a crystal and should run your program with the RCSLOW built-in clock.
Sounds like a really lovely present. You might consider a piezoelectric speaker and have it optionally play tunes.
Post Edited (Mike Green) : 11/21/2009 9:32:47 PM GMT
As far as battery life goes, I always power the LEDs near their peak efficiency and control the intensity by flickering.
I guess I'll just suck it up and use the transistors for high voltage lights.
Thanks again for the help.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
I wonder if this wire is hot...
-Phil
It's not clear to me how the low current condition could manifest under that circumstance.
Considering that the LDO voltage regulator will drop Vcc under the 2.7V brownout threshold when Vdd is around 3.7V, the lowest possible current is around 11 mA.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
I wonder if this wire is hot...
I'm a little hazy here since I haven't used a transistor in awhile, but wouldn't the LED's protective resistor fulfill that requirement?
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
I wonder if this wire is hot...
The problem that you run into with the LED running off 6V is "what happens with the voltage on the I/O pin when the LED is not on?"
An LED is not a perfect diode. If the Propeller I/O pin is set to input mode, there's only a leakage current from the Propeller pin through the imperfect LED to the 6V supply. If the forward voltage of the LED at that current is over 2.1V or so, you're safe since the Propeller pin will be at 3.9V. If the forward voltage of the LED is at 2.1V or less, the protective diode will start to conduct to the Vcc line. If the LED can leak at least 0.5mA without the forward voltage of the LED rising above 2.1V, the protective diode will become damaged and may conduct enough current into the Propeller chip's internal Vcc power bus so that the voltage from point to point across the chip will exceed 4V and damage some of the gate oxide on the chip (which can only handle 4V before damage can occur).
There are a lot of "if"s in the previous paragraph and they can be checked. It may be that you're safe, but it's impossible to be sure without checking. Remember that the voltage drop across a 330 Ohm resistor is low when the current through the resistor is low. I don't know what the "leakage" current through the LED is. It's not one of those things that's carefully characterized on the LED datasheet.
Post Edited (Mike Green) : 11/21/2009 10:29:21 PM GMT
That last post by Mike raised a problem I hadn't considered.
4V was an arbitrary number I threw out for the sake of making the numbers easy, it's actually closer to 3.6V for maximum efficiency but they will run on less.
Suppose the actual shut off voltage is 2.4V.
By Mike's numbers above the Propellor will be safe but there is another problem.
The light will never shut off!
Even if I pull the pin high to 3.3V there will still be a 2.7V differential across the light.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
I wonder if this wire is hot...
By the way, it's getting to be close to Christmas time. There are still lots of SX chips available. If you have an existing design, you could certainly produce this years' batch using the SX (and maybe the next year or two's batches as well). You could spend next year working on the new design.
Certainly the SX will provide a fallback if I can't get up to speed fast enough, but the Propeller assembly code is much more powerful and it has substantially more RAM that's better organized.
The PII looks like it will open up some new doors as well.
I have some Cypress L2 cache lying around that needs 52 pins just for the address and data.
It can crank out a 32 bit DDR bus at 166 MHz, so it could communicate to each cog in turn and provide the equivalent of a 256 bit bus in one cycle of the hub.
That kind of data flow and single instruction per cycle pipelining opens up the possibility of putting together reasonable demos of chaos theory of computational fluid dynamics.
Of course I could make such a demo on my PC, but where is the fun in that?
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
I wonder if this wire is hot...