'DIY opto-isolator furrgetaboutit'
>I tried building a do-it-yourself opto-isolator using a T13/4 red and
> green LED pointing at each other.
There are LEDs and LEDs.
What luminous efficiency did they have?
Modern LEDs are getting around 100+ l/W for white and above and below that
for discrete colours depending on where they are in the spectrum (as lumens
uses human eye response as part of its definition). mW optical / mW
electrical is probabnly a better measure. 100+ lumen/Watt for white is 30%+
Mayhaps a blue LED of suitably good efficiency may be useful as their quanta
have bigger muscles. Dunno how much it matters.
A silicon "solar cell" or several would probably serve you better.
If you have a N cell amorphous panel and are adept at glass cutting you can
cut N cell slivers off the end and have ~ N/2 volt uWatt cells.
Lets say you get 5% efficiency.
Totally converting an eg 20 mA modern white LED you'd get.
Pin ~= 3.2V x 20 mA = 66 mW
Pout = say 15 mW optical.
5% conversion = 750 uW.
Area needed for a cell good for 1000 W/m^2 insolation is about
66E-3/1000 = 66 E-6 m^3 = 66 mm^2.
So about 1 cm^2 of N cell amorphous silicon solar cell illuminated with all
the light from a 20 mA white LED of good efficincy MIGHT be expected to give
you about 500 uW of power or say 100 uA at 5V.
Would be interesting to try.
A small but somewhat larger PV panel illuminated with one LED should give
Get the insolation (InLEDation?) too too low and Vout falls. Amorphous is
somewhat better than crystalline usually in this regard (but lower
A 20mA white at 50+ l/W and about 15 degrees is in the 15 - 25 candella
(Candella is a terrible measure for working out how efficient the LED is as
its ~= a function of 1/(half power double angle)^2 and also of beam shape
which can change naughtily between LEDs of same alleged half power angle.
> > Maybe I should have used a 500M ohm resistor.
IN4007 reverse biased make useful centamegohm resistors.
Something in my brain said you were trying to power the gate, but I see you
are only trying to switch it.
My 'solutions' would be a wee bot of overkill for that.
> There are LEDs and LEDs.
Maybe inserting an OpAmp between the LED and the gate.
On 20 Nov 2008 at 22:15, apptech wrote:
Michael Algernon wrote:
> I tried building a do-it-yourself opto-isolator
> using a T13/4 red and
> green LED pointing at each other.
That doesn't sound like a good idea. I think the reception wavelengths of
LEDs are usually close to their transmission wavelengths.
> I flattened the faces of the LEDs and polished
> them and glued them together with clear glue ( very thin coat ). I
> drove the green LED with about 10ma of current. I reversed biased the
> red LED with a 50M ohm pull-up resistor. I fed the connection between
> the LED and the pull-up resistor into a 74C14 hex inverter ( with
> schmitt triggers ).
Sounds like a long way to go to build half of a 75 cent off the shelf part
> The 74C14 is specified to have .005 ua of input
> current or 5*10**-9 amp.
Learn to say "nano". It means 10**-9.
> The pull-up resistor was supplying about
> 10**-7 amp. So I thought it might work. No it did not. Only when I
> shone a bright flash-lite on it, did it change state.
This means the leakage current wasn't getting you, but not enough of the
right wavelength light was getting into the receiver LED.
> Maybe I should have used a 500M ohm resistor.
Where are you going to get that from? And even if you did, how are you
going to keep it and the board its on and all the other parts super clean to
Personally I think
http://www.mouser.com/Search/ProductDetail.aspx?qs=bfRUmXT2lZgQAJSbX58%2fqw%3d%3d is rather less hassle.
Embed Inc, Littleton Massachusetts, http://www.embedinc.com/products
(978) 742-9014. Gold level PIC consultants since 2000.
> On Nov 20, 2008, at 8:43 AM, Olin Lathrop wrote:
> Michael Algernon wrote:
>> I tried building a do-it-yourself opto-isolator
>> using a T13/4 red and
>> green LED pointing at each other.
> That doesn't sound like a good idea. I think the reception
> wavelengths of
> LEDs are usually close to their transmission wavelengths.
I tried red on red. No joy. Sounds decadent.
>> The 74C14 is specified to have .005 ua of input
>> current or 5*10**-9 amp.
> Learn to say "nano". It means 10**-9.
and 10**-10 is nano-lite.
This is about fun...... not practicality.
Denver, CO 720 222 1309
" dent the UNIVERSE "
All ideas, text, drawings and audio , that are originated by WFT
Electronics ( and it's principals ), that are included with this
signature text are to be deemed to be released to the public domain as
of the date of this communication .
More... (looser matching)
- Last day of these posts
- In 2008
, 2009 only
- New search...