DIY LED Tester

[circuitmangler]

Hi everyone!

I'm sure most people are familiar with the hand-held LED
tester boxes that have come out on the market recently,
e.g. the ones that look like this:

http://www.ledtronics.com/ds/TST-1001/

For a while now I've contemplated how easy it could be to
make one yourself with an emphasis on doing it cheaply.
I've come up with a design, and I'd appreciate any feedback
you may have.

Version 1 - the el cheapo version.

This version just uses a single op amp but it can only test
one led at a time. The schematic can be found here:

http://www.headphoneamp.co.kr/ftp/sijosae/Gallery/Circuit/Circuit-11.gif

To test the led at different currents, just connect it
between the op amp output and different resistor values.

Version 2 - The multiple LED version.

This version just uses one op amp per LED circuit, so to
duplicate the LEDtronic's device I'd need 11 op amps for
the top row and some more for the bi-color LED section.
Alternatively, I could use voltage regulators (such as a
LM317, etc.) to create constant current sources.

So -- it seems I need to find a cheap source of op amps
and/or voltage regulators that can source, say, 50 mA.
So here are the questions:

- Will the LM324N work for this? It's a quad-packaged
op amp that, if I read the datasheet correctly, can
source +/- 60mA (is that per op amp?). Also, I can get
them at glichbuster.com for $0.27 + $1.95 shipping.

- Does anyone know of an alternative op amp/voltage
regulator which is more cost effective?

- Is there another design which could cut down on the
required components?

Thanks in advance for any comments and suggestions!

cm

 

[circuitmangler]

I think I just realized that one op amp can't be used to
test multiple currents -- at least the circuit shown
won't work that way. oh well... back to the drawing board.

cm

 

[itsmoked]

How could you get simpler than the 317? 50 cents a channel in singles.

http://focus.ti.com/lit/ds/symlink/lm317l.pdf

Keith Cress
Flamin Systems, Inc.-

http://www.flaminsystems.com

 

[logbook]

I don't understand the question. If you want to test an LED why not just connect it to a 6V battery through a resistor? You can use different value resistors for different currents and different colours if you like. I can't think of a cheaper test circuit than a resistor!

The reason I chose 6V is that LEDs don't like a lot of reverse voltage so if you plug it in backwards it won't blow up.

 

[itsmoked]

Hi logbook;

The resistor method won't show brightness variations, since each LED's forward voltage characteristic will allow different amounts of current though, even in the same lot of LEDs.

On the other hand who wants to test a bunch of LEDs? lol. If I want to know if one works I just flip my DMM to diode test and check. Usually that lights up functional ones.

Keith Cress
Flamin Systems, Inc.-

http://www.flaminsystems.com

 

[circuitmangler]

itsmoked...

yes, the voltage regulator approach is simple, but the
op amp approach is only $0.09/circuit. The lm317
can supply higher currents, though, and would be useful for
testing LEDs at > 50ma. Having multiple testing circuits
available is nice for doing side-by-side comparisons of
LEDs at different currents (or different LEDs at the same
current.)

Another approach I've seen that could be possibly even
cheaper at the expense of being more work to wire up
is to create a current regulator out of 2 transistors
as in:

http://wolfstone.halloweenhost.com/Lighting/litled_LightEmittingDiodes.html#LEDTester

Yet another possibility is to wire up a FET as a current
regulator, but the cheapest I've seen are 2n7000s for about
$0.13 a piece.

This is mainly an exercise to see how cheaply such a device
could be built from scratch.

 

[OperaHouse]

Don't forget that with regulator constant current method you can test a number of LEDs in series at the same current. This allows you to do a side by side comparison at the same current. You can make it adjustable by placing a 1K pot in parallel with the sense resistor with the tap going to the center leg of the regulator. Suggest a fixed resistor in one leg of the pot to limit the max current to a safe value.

While not as precise, a fixed voltage feeding the base of a transistor will serve as a current source. The emitter resistor will determine the current. This can be adjusted up and down by changing the base voltage. LED is placed in the collector leg. This circuit can be enhanced by driving it with your said op amp. I figure transistors are less than a nickel, your price may vary.

 

[benta]

Don't know where you get the idea that an LM324 can supply 60 mA.
The 60 mA is the maximum short-circuit current, which is a useless parameter for this application.
Source current is guaranteed at 20 mA at 25 C, 10 mA over full temperature range. Sink current 10 resp. 5 mA. Those are the parameters you should design to.

Benta.

 

[IRstuff]

If you use a pair of reasonably matched transistors and a resistor, you can build a Widlar current mirror

TTFN

 

[circuitmangler]

benta - thanks for pointing that out. I looked for limits on
supply current, but couldn't find it on the first datasheet
I found. Now that I have looked at some others, I see where
the supply current limits are explicitly stated. thanks again!

 

[melone]

What are you trying to do, verify that the LED functions, or validate the LED operates with the manufacturers specifications?

 

[logbook]

Personally I can't see the volt drop across LEDs from the same batch being significant compared to a 6V power supply. It's pretty much like Vbe variations in transistors. A 1.2V LED would give a reasonable current accuracy on a 6V rail.

If you want the same current for different types of LEDs (like blue) then a current source is required. Use a PNP transistor emitter into a resistor up to the positive rail. You can chain a load of bases together and use one opamp to drive the bases. Incremental cost is one transistor and one resistor per stage.

 

[circuitmangler]

melone --

the purpose of the these hand-held devices is simply to
provide a convenient way to see what leds look like under
different current loads. they have multiple testing
circuits so you can place them side-by-side for comparison.
So, they could be used for a variety of purposes, i.e.
determining anode/cathode pins, verifing the led works,
determining appropriate driving current, etc. Here are
some more examples of these devices:

http://www.goldmine-elec-products.com/prodinfo.asp?number=G15398

http://jameco.com/webapp/wcs/stores...toreId=10001&catalogId=10001&productId=355805

http://www.allelectronics.com/cgi-bin/item/LT-100/780/LED_TESTER_.html

My purpose is simply to recreate one device as cheaply as
I can with commonly available components.

 

[melone]

Forgive the obvious question, but why not buy an existing unit? Don't forget to factor in your time in the calculation of cost.

 

[IRstuff]

The first item listed only costs $10!!!

Just reading this thread cost more...

Even if you're operating at minimum wage, it'll take you longer than 2 hrs to duplicate all the functionality listed.

TTFN

 

[circuitmangler]

Hi guys,

The point is not to save money but just to gain a better insight into how you could build one yourself.

Thanks for the tips --

cm

 

[IRstuff]

Given that, I would spend the $10 and take it apart for the baseline analysis

TTFN